diff --git a/.gitignore b/.gitignore new file mode 100644 index 000000000..46f3c0430 --- /dev/null +++ b/.gitignore @@ -0,0 +1 @@ +weed diff --git a/.project b/.project index f5551728c..4dee05c73 100644 --- a/.project +++ b/.project @@ -5,11 +5,6 @@ - - com.googlecode.goclipse.goBuilder - - - goclipse.goNature diff --git a/.travis.yml b/.travis.yml index a6f8b8815..79cec6d8a 100644 --- a/.travis.yml +++ b/.travis.yml @@ -1,6 +1,5 @@ language: go go: - - 1.2 - 1.3 - release - tip diff --git a/Dockerfile b/Dockerfile index 1243727a2..06d81ebf0 100644 --- a/Dockerfile +++ b/Dockerfile @@ -1,5 +1,5 @@ FROM cydev/go -RUN go get code.google.com/p/weed-fs/go/weed +RUN go get github.com/chrislusf/weed-fs/go/weed EXPOSE 8080 EXPOSE 9333 VOLUME /data diff --git a/README.md b/README.md index 1808a0277..95d2a5ade 100644 --- a/README.md +++ b/README.md @@ -1,18 +1,335 @@ -weed-fs +Seaweed File System ======= [![Build Status](https://travis-ci.org/chrislusf/weed-fs.svg?branch=master)](https://travis-ci.org/chrislusf/weed-fs) [![GoDoc](https://godoc.org/github.com/chrislusf/weed-fs/go?status.svg)](https://godoc.org/github.com/chrislusf/weed-fs/go) [![RTD](https://readthedocs.org/projects/weed-fs/badge/?version=latest)](http://weed-fs.readthedocs.org/en/latest/) -An official mirror of code.google.com/p/weed-fs. -Moving to github.com to make cooperations easier. -This repo and the google code repo will be kept synchronized. +## Introduction + +Seaweed-FS is a simple and highly scalable distributed file system. There are two objectives: + +1. to store billions of files! +2. to serve the files fast! + +Instead of supporting full POSIX file system semantics, Seaweed-FS choose to implement only a key~file mapping. Similar to the word "NoSQL", you can call it as "NoFS". + +Instead of managing all file metadata in a central master, Seaweed-FS choose to manages file volumes in the central master, and let volume servers manage files and the metadata. This relieves concurrency pressure from the central master and spreads file metadata into volume servers' memories, allowing faster file access with just one disk read operation! + +Seaweed-FS models after [Facebook's Haystack design paper](http://www.usenix.org/event/osdi10/tech/full_papers/Beaver.pdf). + +Seaweed-FS costs only 40 bytes disk storage for each file's metadata. It is so simple with O(1) disk read that you are welcome to challenge the performance with your actual use cases. + + +![](https://api.bintray.com/packages/chrislusf/Weed-FS/seaweed/images/download.png) +https://bintray.com/chrislusf/Weed-FS/seaweed Download latest compiled binaries for different platforms here. + +## Additional Features +* Can choose no replication or different replication level, rack and data center aware +* Automatic master servers failover. No single point of failure, SPOF. +* Automatic Gzip compression depending on file mime type +* Automatic compaction to reclaimed disk spaces after deletion or update +* Servers in the same cluster can have different disk spaces, different file systems, different OS +* Adding/Removing servers do not cause any data re-balancing +* Optional [filer server](https://code.google.com/p/weed-fs/wiki/DirectoriesAndFiles) provides "normal" directories and files via http +* For jpeg pictures, optionally fix the orientation. +* Support Etag, Accept-Range, Last-Modified, etc. + +## Example Usage +By default, the master node runs on port 9333, and the volume nodes runs on port 8080. +Here I will start one master node, and two volume nodes on port 8080 and 8081. Ideally, they should be started from different machines. Here I just use localhost as example. + +Seaweed-FS uses HTTP REST operations to write, read, delete. The return results are JSON or JSONP format. + +### Start Master Server + +``` +> ./weed master + +``` + +### Start Volume Servers ### + +``` +> weed volume -dir="/tmp/data1" -max=5 -mserver="localhost:9333" -port=8080 & +> weed volume -dir="/tmp/data2" -max=10 -mserver="localhost:9333" -port=8081 & + +``` + +### Write File ### +Here is a simple usage on how to save a file: + +``` +> curl http://localhost:9333/dir/assign +{"count":1,"fid":"3,01637037d6","url":"127.0.0.1:8080","publicUrl":"localhost:8080"} +``` +First, send a HTTP request to get an fid and a volume server url. + +``` +> curl -F file=@/home/chris/myphoto.jpg http://127.0.0.1:8080/3,01637037d6 +{"size": 43234} +``` +Second, send a HTTP multipart POST request to the volume server url+'/'+fid, to really store the file content. + +For update, send another POST request with updated file content. + +For deletion, send a http DELETE request +``` +> curl -X DELETE http://127.0.0.1:8080/3,01637037d6 +``` +### Save File Id ### +Now you can save the fid, 3,01637037d6 in this case, to some database field. + +The number 3 here, is a volume id. After the comma, it's one file key, 01, and a file cookie, 637037d6. + +The volume id is an unsigned 32 bit integer. The file key is an unsigned 64bit integer. The file cookie is an unsigned 32bit integer, used to prevent URL guessing. + +The file key and file cookie are both coded in hex. You can store the tuple in your own format, or simply store the fid as string. + +If stored as a string, in theory, you would need 8+1+16+8=33 bytes. A char(33) would be enough, if not more than enough, since most usage would not need 2^32 volumes. + +If space is really a concern, you can store the file id in your own format. You would need one 4-byte integer for volume id, 8-byte long number for file key, 4-byte integer for file cookie. So 16 bytes are enough (more than enough). + +### Read File ### +Here is the example on how to render the URL. +``` +> curl http://localhost:9333/dir/lookup?volumeId=3 +{"locations":[{"publicUrl":"localhost:8080","url":"localhost:8080"}]} +``` +First lookup the volume server's URLs by the file's volumeId. However, since usually there are not too many volume servers, and volumes does not move often, you can cache the results most of the time. Depends on the replication type, one volume can have multiple replica locations. Just randomly pick one location to read. + +Now you can take the public url, render the url or directly read from the volume server via url: +``` + http://localhost:8080/3,01637037d6.jpg +``` +Notice we add an file extension ".jpg" here. It's optional and just one way for the client to specify the file content type. + +If you want a nicer URL, you can use one of these alternative URL formats: +``` + http://localhost:8080/3/01637037d6/my_preferred_name.jpg + http://localhost:8080/3/01637037d6.jpg + http://localhost:8080/3,01637037d6.jpg + http://localhost:8080/3/01637037d6 + http://localhost:8080/3,01637037d6 +``` + +### Rack-Aware and Data Center-Aware Replication ### +Seaweed-FS apply the replication strategy on a volume level. So when you are getting a file id, you can specify the replication strategy. For example: +``` +curl http://localhost:9333/dir/assign?replication=001 +``` + +Here is the meaning of the replication parameter +``` +000: no replication +001: replicate once on the same rack +010: replicate once on a different rack, but same data center +100: replicate once on a different data center +200: replicate twice on two different data center +110: replicate once on a different rack, and once on a different data center +``` + +More details about replication can be found here: +https://code.google.com/p/weed-fs/wiki/RackDataCenterAwareReplication + +You can also set the default replication strategy when starting the master server. + +### Allocate File Key on specific data center ### +Volume servers can start with a specific data center name. +``` + weed volume -dir=/tmp/1 -port=8080 -dataCenter=dc1 + weed volume -dir=/tmp/2 -port=8081 -dataCenter=dc2 +``` +Or the master server can determine the data center via volume server's IP address and settings in weed.conf file. + +Now when requesting a file key, an optional "dataCenter" parameter can limit the assigned volume to the specific data center. For example, this specify +``` + http://localhost:9333/dir/assign?dataCenter=dc1 +``` + +### Other Features ### + * [No Single Point of Failure](https://code.google.com/p/weed-fs/wiki/FailoverMasterServer) + * [Insert with your own keys](https://code.google.com/p/weed-fs/wiki/Optimization#Insert_with_your_own_keys) + * [ Chunking large files](https://code.google.com/p/weed-fs/wiki/Optimization#Upload_large_files) + * [Collection as a Simple Name Space](https://code.google.com/p/weed-fs/wiki/Optimization#Collection_as_a_Simple_Name_Space) + +## Architecture ## +Usually distributed file system split each file into chunks, and a central master keeps a mapping of a filename and a chunk index to chunk handles, and also which chunks each chunk server has. + +This has the draw back that the central master can not handle many small files efficiently, and since all read requests need to go through the chunk master, responses would be slow for many concurrent web users. + +Instead of managing chunks, Seaweed-FS choose to manage data volumes in the master server. Each data volume is size 32GB, and can hold a lot of files. And each storage node can has many data volumes. So the master node only needs to store the metadata about the volumes, which is fairly small amount of data and pretty static most of the time. + +The actual file metadata is stored in each volume on volume servers. Since each volume server only manage metadata of files on its own disk, and only 16 bytes for each file, all file access can read file metadata just from memory and only needs one disk operation to actually read file data. + +For comparison, consider that an xfs inode structure in Linux is 536 bytes. + +### Master Server and Volume Server ### +The architecture is fairly simple. The actual data is stored in volumes on storage nodes. One volume server can have multiple volumes, and can both support read and write access with basic authentication. + +All volumes are managed by a master server. The master server contains volume id to volume server mapping. This is fairly static information, and could be cached easily. + +On each write request, the master server also generates a file key, which is a growing 64bit unsigned integer. Since the write requests are not as busy as read requests, one master server should be able to handle the concurrency well. + + +### Write and Read files ### + +When a client sends a write request, the master server returns for the file. The client then contact the volume node and POST the file content via REST. + +When a client needs to read a file based on , it can ask the master server by the for the , or from cache. Then the client can HTTP GET the content via REST, or just render the URL on web pages and let browsers to fetch the content. + +Please see the example for details on write-read process. + +### Storage Size ### +In current implementation, each volume can be size of 8x2^32^=32G bytes. This is because of aligning contents to 8 bytes. We can be easily increased to 64G, or 128G, or more, by changing 2 lines of code, at the cost of some wasted padding space due to alignment. + +There can be 2^32^ volumes. So total system size is 8 x 2^32^ x 2^32^ = 8 x 4G x 4G = 128GG bytes. (Sorry, I don't know the word for giga of giga bytes.) + +Each individual file size is limited to the volume size. + +### Saving memory ### +All file meta information on volume server is readable from memory without disk access. Each file just takes an 16-byte map entry of <64bit key, 32bit offset, 32bit size>. Of course, each map entry has its own the space cost for the map. But usually the disk runs out before the memory does. + + +## Compared to Other File Systems## +Frankly, I don't use other distributed file systems too often. All seems more complicated than necessary. Please correct me if anything here is wrong. + +### Compared to Ceph ### +Ceph can be setup similar to Seaweed-FS as a key~blob store. It is much more complicated, with the need to support layers on top of it. Here is a more detailed comparison. https://code.google.com/p/weed-fs/issues/detail?id=44 + +Seaweed-FS is meant to be fast and simple, both during usage and during setup. If you do not understand how it works when you reach here, we failed! Jokes aside, you should not need any consulting service for it. + +Seaweed-FS has a centralized master to lookup free volumes, while Ceph uses hashing to locate its objects. Having a centralized master makes it easy to code and manage. HDFS/GFS has the single name node for years. Seaweed-FS now support multiple master nodes. + +Ceph hashing avoids SPOF, but makes it complicated when moving or adding servers. + +### Compared to HDFS ### +HDFS uses the chunk approach for each file, and is ideal for streaming large files. + +Seaweed-FS is ideal for serving relatively smaller files quickly and concurrently. + +Seaweed-FS can also store extra large files by splitting them into manageable data chunks, and store the file ids of the data chunks into a meta chunk. This is managed by "weed upload/download" tool, and the weed master or volume servers are agnostic about it. + +### Compared to MogileFS### +Seaweed-FS has 2 components: directory server, storage nodes. + +MogileFS has 3 components: tracers, database, storage nodes. + +One more layer means slower access, more operation complexity, more failure possibility. + +### Compared to GlusterFS ### +Seaweed-FS is not POSIX compliant, and has simple implementation. + +GlusterFS is POSIX compliant, much more complex. + +### Compared to Mongo's GridFS ### +Mongo's GridFS splits files into chunks and manage chunks in the central mongodb. For every read or write request, the database needs to query the metadata. It's OK if this is not a bottleneck yet, but for a lot of concurrent reads this unnecessary query could slow things down. + +Since files are chunked(default to 256KB), there will be multiple metadata readings and multiple chunk readings, linear to the file size. One 2.56MB file would require at least 20 disk read requests. + +On the contrary, Seaweed-FS uses large file volume of 32G size to store lots of files, and only manages file volumes in the master server. Each volume manages file metadata themselves. So all the file metadata is spread onto the volume nodes memories, and just one disk read is needed. + +## Dev plan ## + +More tools and documentation, on how to maintain and scale the system. For example, how to move volumes, automatically balancing data, how to grow volumes, how to check system status, etc. + +This is a super exciting project! And I need helpers! + + +## Installation guide for users who are not familiar with golang + +step 1: install go on your machine and setup the environment by following the instructions from the following link: + +https://golang.org/doc/install + +make sure you set up your $GOPATH + + +step 2: also you may need to install Mercurial by following the instructions below + +http://mercurial.selenic.com/downloads + + +step 3: download, compile, and install the project by executing the following command + +go get github.com/chrislusf/weed-fs/go/weed + +once this is done, you should see the executable "weed" under $GOPATH/bin + +step 4: after you modify your code locally, you could start a local build by calling "go install" under $GOPATH/src/github.com/chrislusf/weed-fs/go/weed + +## Reference -For documents and bug reporting, Please visit - http://weed-fs.googlecode.com - For pre-compiled releases, - https://bintray.com/chrislusf/Weed-FS/weed + https://bintray.com/chrislusf/Weed-FS/seaweed + +## Disk Related topics ## + +### Hard Drive Performance ### +When testing read performance on Seaweed-FS, it basically becomes performance test your hard drive's random read speed. Hard Drive usually get 100MB/s~200MB/s. + +### Solid State Disk + +To modify or delete small files, SSD must delete a whole block at a time, and move content in existing blocks to a new block. SSD is fast when brand new, but will get fragmented over time and you have to garbage collect, compacting blocks. Seaweed-FS is friendly to SSD since it is append-only. Deletion and compaction are done on volume level in the background, not slowing reading and not causing fragmentation. + +## Not Planned + +POSIX support + + +## Benchmark + +My Own Unscientific Single Machine Results on Mac Book with Solid State Disk, CPU: 1 Intel Core i7 2.2GHz. + +Write 1 million 1KB file: +``` +Concurrency Level: 64 +Time taken for tests: 182.456 seconds +Complete requests: 1048576 +Failed requests: 0 +Total transferred: 1073741824 bytes +Requests per second: 5747.01 [#/sec] +Transfer rate: 5747.01 [Kbytes/sec] + +Connection Times (ms) + min avg max std +Total: 0.3 10.9 430.9 5.7 + +Percentage of the requests served within a certain time (ms) + 50% 10.2 ms + 66% 12.0 ms + 75% 12.6 ms + 80% 12.9 ms + 90% 14.0 ms + 95% 14.9 ms + 98% 16.2 ms + 99% 17.3 ms + 100% 430.9 ms +``` + +Randomly read 1 million files: +``` +Concurrency Level: 64 +Time taken for tests: 80.732 seconds +Complete requests: 1048576 +Failed requests: 0 +Total transferred: 1073741824 bytes +Requests per second: 12988.37 [#/sec] +Transfer rate: 12988.37 [Kbytes/sec] + +Connection Times (ms) + min avg max std +Total: 0.0 4.7 254.3 6.3 +Percentage of the requests served within a certain time (ms) + 50% 2.6 ms + 66% 2.9 ms + 75% 3.7 ms + 80% 4.7 ms + 90% 10.3 ms + 95% 16.6 ms + 98% 26.3 ms + 99% 34.8 ms + 100% 254.3 ms +``` diff --git a/docs/api.rst b/docs/api.rst index 4f3806be9..b67463960 100644 --- a/docs/api.rst +++ b/docs/api.rst @@ -69,17 +69,6 @@ One volume servers one write a time. If you need to increase concurrency, you ca This generates 4 empty volumes. -Upload File Directly -*********************************** - -.. code-block:: bash - - curl -F file=@/home/chris/myphoto.jpg http://localhost:9333/submit - {"fid":"3,01fbe0dc6f1f38","fileName":"myphoto.jpg", - "fileUrl":"localhost:8080/3,01fbe0dc6f1f38","size":68231} - -This API is a little convenient. The master server would contact itself via HTTP to get an file id and store it to the right volume server. It is a convenient API and does not support different parameters when assigning file id. - Check System Status *********************************** @@ -191,17 +180,18 @@ Upload File curl -F file=@/home/chris/myphoto.jpg http://127.0.0.1:8080/3,01637037d6 {"size": 43234} -The size returned is the size stored on WeedFS, sometimes the file is automatically gzipped based on the mime type. +The size returned is the size stored on Seaweed-FS, sometimes the file is automatically gzipped based on the mime type. Upload File Directly *********************************** .. code-block:: bash - curl -F file=@/home/chris/myphoto.jpg http://localhost:8080/submit + curl -F file=@/home/chris/myphoto.jpg http://localhost:9333/submit {"fid":"3,01fbe0dc6f1f38","fileName":"myphoto.jpg","fileUrl":"localhost:8080/3,01fbe0dc6f1f38","size":68231} -This API is a little convenient. The volume server would contact the master to get an file id and store it to the right volume server(not necessarily itself). +This API is just for convenience. The master server would get an file id and store the file to the right volume server. +It is a convenient API and does not support different parameters when assigning file id. (or you can add the support and send a push request.) Delete File *********************************** diff --git a/docs/benchmarks.rst b/docs/benchmarks.rst index e01e6e4ba..d0b71428f 100644 --- a/docs/benchmarks.rst +++ b/docs/benchmarks.rst @@ -1,7 +1,7 @@ Benchmarks ====================== -Do we really need the benchmark? People always use benchmark to compare systems. But benchmarks are misleading. The resources, e.g., CPU, disk, memory, network, all matter a lot. And with Weed File System, single node vs multiple nodes, benchmarking on one machine vs several multiple machines, all matter a lot. +Do we really need the benchmark? People always use benchmark to compare systems. But benchmarks are misleading. The resources, e.g., CPU, disk, memory, network, all matter a lot. And with Seaweed File System, single node vs multiple nodes, benchmarking on one machine vs several multiple machines, all matter a lot. Here is the steps on how to run benchmark if you really need some numbers. @@ -38,7 +38,22 @@ Many options are options are configurable. Please check the help content: Common Problems ############################### -The most common problem is "too many open files" error. This is because the test itself starts too many network connections on one single machine. In my local macbook, if I ran "random read" following writing right away, the error happens always. I have to run "weed benchmark -write=false" to run the reading test only. Also, changing the concurrency level to "-c=16" would also help. +The most common +I start weed servers in one console for simplicity. Better run servers on different consoles. + +For more realistic tests, please start them on different machines. + +.. code-block:: bash + + # prepare directories + mkdir 3 4 5 + # start 3 servers + ./weed server -dir=./3 -master.port=9333 -volume.port=8083 & + ./weed volume -dir=./4 -port=8084 & + ./weed volume -dir=./5 -port=8085 & + ./weed benchmark -server=localhost:9333 + + problem is "too many open files" error. This is because the test itself starts too many network connections on one single machine. In my local macbook, if I ran "random read" following writing right away, the error happens always. I have to run "weed benchmark -write=false" to run the reading test only. Also, changing the concurrency level to "-c=16" would also help. My own unscientific single machine results ################################################### diff --git a/docs/clients.rst b/docs/clients.rst index 66f7bca5f..6c6f09729 100644 --- a/docs/clients.rst +++ b/docs/clients.rst @@ -8,17 +8,17 @@ Clients +---------------------------------------------------------------------------------+--------------+-----------+ | `WeedPHP `_ | Mic Johnson | PHP | +---------------------------------------------------------------------------------+--------------+-----------+ -| `Weed-FS Symfony bundle `_ | Mic Johnson | PHP | +| `Seaweed-FS Symfony bundle `_ | Mic Johnson | PHP | +---------------------------------------------------------------------------------+--------------+-----------+ -| `Weed-FS Node.js client `_ | Aaron Blakely| Javascript| +| `Seaweed-FS Node.js client `_ | Aaron Blakely| Javascript| +---------------------------------------------------------------------------------+--------------+-----------+ -| `Amazon S3 API for weed-fs `_ | Tamás Gulácsi| Go | +| `Amazon S3 API for Seaweed-FS `_ | Tamás Gulácsi| Go | +---------------------------------------------------------------------------------+--------------+-----------+ | `File store upload test `_ | Tamás Gulácsi| Go | +---------------------------------------------------------------------------------+--------------+-----------+ -| `Java Weed-FS client `_ | Xu Zhang | Java | +| `Java Seaweed-FS client `_ | Xu Zhang | Java | +---------------------------------------------------------------------------------+--------------+-----------+ -| `Java Weed-FS client 2 `_ | Zenria | Java | +| `Java Seaweed-FS client 2 `_ | Zenria | Java | +---------------------------------------------------------------------------------+--------------+-----------+ | `Python-weed `_ | Darkdarkfruit| Python | +---------------------------------------------------------------------------------+--------------+-----------+ @@ -26,7 +26,7 @@ Clients +---------------------------------------------------------------------------------+--------------+-----------+ | `Camlistore blobserver Storage `_ | Tamás Gulácsi| Go | +---------------------------------------------------------------------------------+--------------+-----------+ -| `Scala Weed-FS client `_ | Chiradip | Scala | +| `Scala Seaweed-FS client `_ | Chiradip | Scala | +---------------------------------------------------------------------------------+--------------+-----------+ | `Module for kohana `_ | Bububa | PHP | +---------------------------------------------------------------------------------+--------------+-----------+ @@ -35,10 +35,10 @@ Clients | `Django-weed `_ | ProstoKSI | Python | +---------------------------------------------------------------------------------+--------------+-----------+ -Projects using Weed-Fs +Projects using Seaweed-FS ################################### -* An `email River Plugin `_ for Elasticsearch uses weed-fs server to save attachments +* An `email River Plugin `_ for Elasticsearch uses Seaweed-FS server to save attachments -Websites using Weed-Fs +Websites using Seaweed-FS ################################### -* `Email to create Web Pages `_ uses weed-fs to save email attachments. \ No newline at end of file +* `Email to create Web Pages `_ uses Seaweed-FS to save email attachments. \ No newline at end of file diff --git a/docs/directories.rst b/docs/directories.rst index 57d5d246d..aa14da4f7 100644 --- a/docs/directories.rst +++ b/docs/directories.rst @@ -1,7 +1,7 @@ Directories and files =========================== -When talking about file systems, many people would assume directories, list files under a directory, etc. These are expected if we want to hook up Weed File System with linux by FUSE, or with Hadoop, etc. +When talking about file systems, many people would assume directories, list files under a directory, etc. These are expected if we want to hook up Seaweed File System with linux by FUSE, or with Hadoop, etc. Sample usage ##################### @@ -37,7 +37,7 @@ Design A common file system would use inode to store meta data for each folder and file. The folder tree structure are usually linked. And sub folders and files are usually organized as an on-disk b+tree or similar variations. This scales well in terms of storage, but not well for fast file retrieval due to multiple disk access just for the file meta data, before even trying to get the file content. -WeedFS wants to make as small number of disk access as possible, yet still be able to store a lot of file metadata. So we need to think very differently. +Seaweed-FS wants to make as small number of disk access as possible, yet still be able to store a lot of file metadata. So we need to think very differently. From a full file path to get to the file content, there are several steps: @@ -48,7 +48,7 @@ From a full file path to get to the file content, there are several steps: file_id => data_block -Because default WeedFS only provides file_id=>data_block mapping, the first 2 steps need to be implemented. +Because default Seaweed-FS only provides file_id=>data_block mapping, the first 2 steps need to be implemented. There are several data features I noticed: @@ -122,7 +122,7 @@ The LevelDB implementation may be switched underneath to external data storage, Also, a HA feature will be added, so that multiple "weed filer" instance can share the same set of view of files. -Later, FUSE or HCFS plugins will be created, to really integrate WeedFS to existing systems. +Later, FUSE or HCFS plugins will be created, to really integrate Seaweed-FS to existing systems. Helps Wanted ######################## diff --git a/docs/failover.rst b/docs/failover.rst index f244918d8..506a33e3b 100644 --- a/docs/failover.rst +++ b/docs/failover.rst @@ -7,7 +7,7 @@ Introduction Some user will ask for no single point of failure. Although google runs its file system with a single master for years, no SPOF seems becoming a criteria for architects to pick solutions. -Luckily, it's not too difficult to enable Weed File System with failover master servers. +Luckily, it's not too difficult to enable Seaweed File System with failover master servers. Cheat Sheet: Startup multiple servers ######################################## diff --git a/docs/gettingstarted.rst b/docs/gettingstarted.rst index 0b598a2bf..ad28319fe 100644 --- a/docs/gettingstarted.rst +++ b/docs/gettingstarted.rst @@ -1,9 +1,9 @@ Getting started =================================== -Installing Weed-Fs +Installing Seaweed-FS ################################### -Download a proper version from `WeedFS download page `_. +Download a proper version from `Seaweed-FS download page `_. Decompress the downloaded file. You will only find one executable file, either "weed" on most systems or "weed.exe" on windows. @@ -57,7 +57,7 @@ Actually, forget about previous commands. You can setup one master server and on # use "weed server -h" to find out more ./weed server -master.port=9333 -volume.port=8080 -dir="./data" -Testing Weed-Fs +Testing Seaweed-FS ################################### With the master and volume server up, now what? Let's pump in a lot of files into the system! @@ -77,7 +77,7 @@ Then, you can simply check "du -m -s /some/big/folder" to see the actual disk us Now you can use your tools to hit weed-fs as hard as you can. -Using Weed-Fs in docker +Using Seaweed-FS in docker #################################### You can use image "cydev/weed" or build your own with `dockerfile `_ in the root of repo. diff --git a/docs/optimization.rst b/docs/optimization.rst index 49f0656c8..9d96de060 100644 --- a/docs/optimization.rst +++ b/docs/optimization.rst @@ -1,14 +1,14 @@ Optimization ============== -Here are the strategies or best ways to optimize WeedFS. +Here are the strategies or best ways to optimize Seaweed-FS. Increase concurrent writes ################################ -By default, WeedFS grows the volumes automatically. For example, for no-replication volumes, there will be concurrently 7 writable volumes allocated. +By default, Seaweed-FS grows the volumes automatically. For example, for no-replication volumes, there will be concurrently 7 writable volumes allocated. -If you want to distribute writes to more volumes, you can do so by instructing WeedFS master via this URL. +If you want to distribute writes to more volumes, you can do so by instructing Seaweed-FS master via this URL. .. code-block:: bash @@ -31,14 +31,14 @@ More hard drives will give you better write/read throughput. Gzip content ################################ -WeedFS determines the file can be gzipped based on the file name extension. So if you submit a textual file, it's better to use an common file name extension, like ".txt", ".html", ".js", ".css", etc. If the name is unknown, like ".go", WeedFS will not gzip the content, but just save the content as is. +Seaweed-FS determines the file can be gzipped based on the file name extension. So if you submit a textual file, it's better to use an common file name extension, like ".txt", ".html", ".js", ".css", etc. If the name is unknown, like ".go", Seaweed-FS will not gzip the content, but just save the content as is. -You can also manually gzip content before submission. If you do so, make sure the submitted file has file name with ends with ".gz". For example, "my.css" can be gzipped to "my.css.gz" and sent to WeedFS. When retrieving the content, if the http client supports "gzip" encoding, the gzipped content would be sent back. Otherwise, the unzipped content would be sent back. +You can also manually gzip content before submission. If you do so, make sure the submitted file has file name with ends with ".gz". For example, "my.css" can be gzipped to "my.css.gz" and sent to Seaweed-FS. When retrieving the content, if the http client supports "gzip" encoding, the gzipped content would be sent back. Otherwise, the unzipped content would be sent back. Memory consumption ################################# -For volume servers, the memory consumption is tightly related to the number of files. For example, one 32G volume can easily have 1.5 million files if each file is only 20KB. To store the 1.5 million entries of meta data in memory, currently WeedFS consumes 36MB memory, about 24bytes per entry in memory. So if you allocate 64 volumes(2TB), you would need 2~3GB memory. However, if the average file size is larger, say 200KB, only 200~300MB memory is needed. +For volume servers, the memory consumption is tightly related to the number of files. For example, one 32G volume can easily have 1.5 million files if each file is only 20KB. To store the 1.5 million entries of meta data in memory, currently Seaweed-FS consumes 36MB memory, about 24bytes per entry in memory. So if you allocate 64 volumes(2TB), you would need 2~3GB memory. However, if the average file size is larger, say 200KB, only 200~300MB memory is needed. Theoretically the memory consumption can go even lower by compacting since the file ids are mostly monotonically increasing. I did not invest time on that yet since the memory consumption, 24bytes/entry(including uncompressed 8bytes file id, 4 bytes file size, plus additional map data structure cost) is already pretty low. But I welcome any one to compact these data in memory even more efficiently. @@ -106,7 +106,7 @@ In case you need to delete them later, you can go to the volume servers and dele Logging ############################## -When going to production, you will want to collect the logs. WeedFS uses glog. Here are some examples: +When going to production, you will want to collect the logs. Seaweed-FS uses glog. Here are some examples: .. code-block:: bash diff --git a/docs/replication.rst b/docs/replication.rst index 136f41b97..54e4b2503 100644 --- a/docs/replication.rst +++ b/docs/replication.rst @@ -1,6 +1,6 @@ Replication =================================== -Weed-FS can support replication. The replication is implemented not on file level, but on volume level. +Seaweed-FS can support replication. The replication is implemented not on file level, but on volume level. How to use ################################### @@ -58,7 +58,7 @@ Each replication type will physically create x+y+z+1 copies of volume data files Example topology configuration ################################### -The WeedFS master server tries to read the default topology configuration file are read from /etc/weedfs/weedfs.conf, if it exists. The topology setting to configure data center and racks file format is as this. +The Seaweed-FS master server tries to read the default topology configuration file are read from /etc/weedfs/weedfs.conf, if it exists. The topology setting to configure data center and racks file format is as this. .. code-block:: xml diff --git a/docs/usecases.rst b/docs/usecases.rst index c3cb8e3af..8a9a6d44a 100644 --- a/docs/usecases.rst +++ b/docs/usecases.rst @@ -41,7 +41,7 @@ The wrong way to send it: curl -H "Content-Type:image/png" -F file=@myImage.png http://127.0.0.1:8080/5,2730a7f18b44 -Securing WeedFS +Securing Seaweed-FS ############################# The simple way is to front all master and volume servers with firewall. diff --git a/go/filer/client_operations.go b/go/filer/client_operations.go index 5d89fc865..b38368735 100644 --- a/go/filer/client_operations.go +++ b/go/filer/client_operations.go @@ -1,12 +1,12 @@ package filer -import () - import ( - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "fmt" + + "github.com/chrislusf/weed-fs/go/util" + "net/url" ) diff --git a/go/filer/directory.go b/go/filer/directory.go index 956a2f504..66d1aeba5 100644 --- a/go/filer/directory.go +++ b/go/filer/directory.go @@ -1,7 +1,5 @@ package filer -import () - type DirectoryId int32 type DirectoryEntry struct { diff --git a/go/filer/directory_in_map.go b/go/filer/directory_in_map.go index 9c2ecdf80..ee601066c 100644 --- a/go/filer/directory_in_map.go +++ b/go/filer/directory_in_map.go @@ -2,15 +2,19 @@ package filer import ( "bufio" - "code.google.com/p/weed-fs/go/util" "fmt" "io" "os" "path/filepath" "strconv" "strings" + "sync" + + "github.com/chrislusf/weed-fs/go/util" ) +var writeLock sync.Mutex //serialize changes to dir.log + type DirectoryEntryInMap struct { Name string Parent *DirectoryEntryInMap @@ -25,23 +29,25 @@ type DirectoryManagerInMap struct { isLoading bool } -func (dm *DirectoryManagerInMap) NewDirectoryEntryInMap(parent *DirectoryEntryInMap, name string) (d *DirectoryEntryInMap) { +func (dm *DirectoryManagerInMap) NewDirectoryEntryInMap(parent *DirectoryEntryInMap, name string) (d *DirectoryEntryInMap, err error) { + writeLock.Lock() + defer writeLock.Unlock() d = &DirectoryEntryInMap{Name: name, Parent: parent, SubDirectories: make(map[string]*DirectoryEntryInMap)} - dm.max++ - d.Id = dm.max parts := make([]string, 0) for p := d; p != nil && p.Name != ""; p = p.Parent { parts = append(parts, p.Name) } n := len(parts) if n <= 0 { - return d + return nil, fmt.Errorf("Failed to create folder %s/%s", parent.Name, name) } for i := 0; i < n/2; i++ { parts[i], parts[n-1-i] = parts[n-1-i], parts[i] } + dm.max++ + d.Id = dm.max dm.log("add", "/"+strings.Join(parts, "/"), strconv.Itoa(int(d.Id))) - return d + return d, nil } func (dm *DirectoryManagerInMap) log(words ...string) { @@ -157,7 +163,11 @@ func (dm *DirectoryManagerInMap) loadDirectory(dirPath string, dirId DirectoryId if i != len(parts)-1 { return fmt.Errorf("%s should be created after parent %s!", dirPath, parts[i]) } - sub = dm.NewDirectoryEntryInMap(dir, parts[i]) + var err error + sub, err = dm.NewDirectoryEntryInMap(dir, parts[i]) + if err != nil { + return err + } if sub.Id != dirId { return fmt.Errorf("%s should be have id %v instead of %v!", dirPath, sub.Id, dirId) } @@ -178,7 +188,11 @@ func (dm *DirectoryManagerInMap) makeDirectory(dirPath string) (dir *DirectoryEn for i := 1; i < len(parts); i++ { sub, ok := dir.SubDirectories[parts[i]] if !ok { - sub = dm.NewDirectoryEntryInMap(dir, parts[i]) + var err error + sub, err = dm.NewDirectoryEntryInMap(dir, parts[i]) + if err != nil { + return nil, false + } dir.SubDirectories[parts[i]] = sub created = true } @@ -193,6 +207,8 @@ func (dm *DirectoryManagerInMap) MakeDirectory(dirPath string) (DirectoryId, err } func (dm *DirectoryManagerInMap) MoveUnderDirectory(oldDirPath string, newParentDirPath string, newName string) error { + writeLock.Lock() + defer writeLock.Unlock() oldDir, oe := dm.findDirectory(oldDirPath) if oe != nil { return oe @@ -223,6 +239,8 @@ func (dm *DirectoryManagerInMap) ListDirectories(dirPath string) (dirNames []Dir return dirNames, nil } func (dm *DirectoryManagerInMap) DeleteDirectory(dirPath string) error { + writeLock.Lock() + defer writeLock.Unlock() if dirPath == "/" { return fmt.Errorf("Can not delete %s", dirPath) } diff --git a/go/filer/filer.go b/go/filer/filer.go index de877fc1f..bf4a1cb19 100644 --- a/go/filer/filer.go +++ b/go/filer/filer.go @@ -1,7 +1,5 @@ package filer -import () - type FileId string //file id on weedfs type FileEntry struct { diff --git a/go/filer/filer_embedded.go b/go/filer/filer_embedded.go index a3b64d37b..0b9b4e668 100644 --- a/go/filer/filer_embedded.go +++ b/go/filer/filer_embedded.go @@ -1,11 +1,12 @@ package filer import ( - "code.google.com/p/weed-fs/go/operation" "errors" "fmt" "path/filepath" "strings" + + "github.com/chrislusf/weed-fs/go/operation" ) type FilerEmbedded struct { diff --git a/go/filer/files_in_leveldb.go b/go/filer/files_in_leveldb.go index dbf2e6c52..781bb0e5f 100644 --- a/go/filer/files_in_leveldb.go +++ b/go/filer/files_in_leveldb.go @@ -2,7 +2,8 @@ package filer import ( "bytes" - "code.google.com/p/weed-fs/go/glog" + + "github.com/chrislusf/weed-fs/go/glog" "github.com/syndtr/goleveldb/leveldb" "github.com/syndtr/goleveldb/leveldb/util" ) diff --git a/go/glog/convenient_api.go b/go/glog/convenient_api.go index 3c378083f..cb43d60e2 100644 --- a/go/glog/convenient_api.go +++ b/go/glog/convenient_api.go @@ -1,7 +1,5 @@ package glog -import () - /* Copying the original glog because it is missing several convenient methods. 1. remove nano time in log format diff --git a/go/images/orientation.go b/go/images/orientation.go index a7c126fc1..4bff89311 100644 --- a/go/images/orientation.go +++ b/go/images/orientation.go @@ -2,11 +2,12 @@ package images import ( "bytes" - "github.com/rwcarlsen/goexif/exif" "image" "image/draw" "image/jpeg" "log" + + "github.com/rwcarlsen/goexif/exif" ) //many code is copied from http://camlistore.org/pkg/images/images.go @@ -21,7 +22,10 @@ func FixJpgOrientation(data []byte) (oriented []byte) { } angle := 0 flipMode := FlipDirection(0) - orient := tag.Int(0) + orient, err := tag.Int(0) + if err != nil { + return data + } switch orient { case topLeftSide: // do nothing diff --git a/go/images/resizing.go b/go/images/resizing.go index 08a1e15d2..e9de5f7d7 100644 --- a/go/images/resizing.go +++ b/go/images/resizing.go @@ -2,11 +2,12 @@ package images import ( "bytes" - "github.com/disintegration/imaging" "image" "image/gif" "image/jpeg" "image/png" + + "github.com/disintegration/imaging" ) func Resized(ext string, data []byte, width, height int) (resized []byte, w int, h int) { diff --git a/go/operation/assign_file_id.go b/go/operation/assign_file_id.go index 018e1d763..672bfa99c 100644 --- a/go/operation/assign_file_id.go +++ b/go/operation/assign_file_id.go @@ -1,12 +1,13 @@ package operation import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "net/url" "strconv" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" ) type AssignResult struct { @@ -17,7 +18,7 @@ type AssignResult struct { Error string `json:"error,omitempty"` } -func Assign(server string, count int, replication string, collection string) (*AssignResult, error) { +func Assign(server string, count int, replication string, collection string, ttl string) (*AssignResult, error) { values := make(url.Values) values.Add("count", strconv.Itoa(count)) if replication != "" { @@ -26,6 +27,9 @@ func Assign(server string, count int, replication string, collection string) (*A if collection != "" { values.Add("collection", collection) } + if ttl != "" { + values.Add("ttl", ttl) + } jsonBlob, err := util.Post("http://"+server+"/dir/assign", values) glog.V(2).Info("assign result :", string(jsonBlob)) if err != nil { diff --git a/go/operation/data_struts.go b/go/operation/data_struts.go index 09fab05d1..4980f9913 100644 --- a/go/operation/data_struts.go +++ b/go/operation/data_struts.go @@ -1,7 +1,5 @@ package operation -import () - type JoinResult struct { VolumeSizeLimit uint64 `json:"VolumeSizeLimit,omitempty"` Error string `json:"error,omitempty"` diff --git a/go/operation/delete_content.go b/go/operation/delete_content.go index e4f9d39bb..416a852b3 100644 --- a/go/operation/delete_content.go +++ b/go/operation/delete_content.go @@ -1,12 +1,13 @@ package operation import ( - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "net/url" "strings" "sync" + + "github.com/chrislusf/weed-fs/go/util" ) type DeleteResult struct { diff --git a/go/operation/list_masters.go b/go/operation/list_masters.go index 491c79f20..7ada94243 100644 --- a/go/operation/list_masters.go +++ b/go/operation/list_masters.go @@ -1,9 +1,10 @@ package operation import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" "encoding/json" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" ) type ClusterStatusResult struct { diff --git a/go/operation/lookup.go b/go/operation/lookup.go index f191bfdbf..e6b6658da 100644 --- a/go/operation/lookup.go +++ b/go/operation/lookup.go @@ -1,7 +1,6 @@ package operation import ( - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" _ "fmt" @@ -9,6 +8,8 @@ import ( "net/url" "strings" "time" + + "github.com/chrislusf/weed-fs/go/util" ) type Location struct { diff --git a/go/operation/submit.go b/go/operation/submit.go index 9191f7d9a..62db46617 100644 --- a/go/operation/submit.go +++ b/go/operation/submit.go @@ -2,13 +2,14 @@ package operation import ( "bytes" - "code.google.com/p/weed-fs/go/glog" "io" "mime" "os" "path" "strconv" "strings" + + "github.com/chrislusf/weed-fs/go/glog" ) type FilePart struct { @@ -20,6 +21,7 @@ type FilePart struct { ModTime int64 //in seconds Replication string Collection string + Ttl string Server string //this comes from assign result Fid string //this comes from assign result, but customizable } @@ -32,12 +34,12 @@ type SubmitResult struct { Error string `json:"error,omitempty"` } -func SubmitFiles(master string, files []FilePart, replication string, collection string, maxMB int) ([]SubmitResult, error) { +func SubmitFiles(master string, files []FilePart, replication string, collection string, ttl string, maxMB int) ([]SubmitResult, error) { results := make([]SubmitResult, len(files)) for index, file := range files { results[index].FileName = file.FileName } - ret, err := Assign(master, len(files), replication, collection) + ret, err := Assign(master, len(files), replication, collection, ttl) if err != nil { for index, _ := range files { results[index].Error = err.Error() @@ -112,7 +114,7 @@ func (fi FilePart) Upload(maxMB int, master string) (retSize uint32, err error) chunks := fi.FileSize/chunkSize + 1 fids := make([]string, 0) for i := int64(0); i < chunks; i++ { - id, count, e := upload_one_chunk(fi.FileName+"-"+strconv.FormatInt(i+1, 10), io.LimitReader(fi.Reader, chunkSize), master, fi.Replication, fi.Collection) + id, count, e := upload_one_chunk(fi.FileName+"-"+strconv.FormatInt(i+1, 10), io.LimitReader(fi.Reader, chunkSize), master, fi.Replication, fi.Collection, fi.Ttl) if e != nil { return 0, e } @@ -130,8 +132,8 @@ func (fi FilePart) Upload(maxMB int, master string) (retSize uint32, err error) return } -func upload_one_chunk(filename string, reader io.Reader, master, replication string, collection string) (fid string, size uint32, e error) { - ret, err := Assign(master, 1, replication, collection) +func upload_one_chunk(filename string, reader io.Reader, master, replication string, collection string, ttl string) (fid string, size uint32, e error) { + ret, err := Assign(master, 1, replication, collection, ttl) if err != nil { return "", 0, err } diff --git a/go/operation/system_message.pb.go b/go/operation/system_message.pb.go index 45ae8a648..9f00dd74d 100644 --- a/go/operation/system_message.pb.go +++ b/go/operation/system_message.pb.go @@ -15,12 +15,10 @@ It has these top-level messages: package operation import proto "code.google.com/p/goprotobuf/proto" -import json "encoding/json" import math "math" -// Reference proto, json, and math imports to suppress error if they are not otherwise used. +// Reference imports to suppress errors if they are not otherwise used. var _ = proto.Marshal -var _ = &json.SyntaxError{} var _ = math.Inf type VolumeInformationMessage struct { @@ -33,6 +31,7 @@ type VolumeInformationMessage struct { ReadOnly *bool `protobuf:"varint,7,opt,name=read_only" json:"read_only,omitempty"` ReplicaPlacement *uint32 `protobuf:"varint,8,req,name=replica_placement" json:"replica_placement,omitempty"` Version *uint32 `protobuf:"varint,9,opt,name=version,def=2" json:"version,omitempty"` + Ttl *uint32 `protobuf:"varint,10,opt,name=ttl" json:"ttl,omitempty"` XXX_unrecognized []byte `json:"-"` } @@ -105,6 +104,13 @@ func (m *VolumeInformationMessage) GetVersion() uint32 { return Default_VolumeInformationMessage_Version } +func (m *VolumeInformationMessage) GetTtl() uint32 { + if m != nil && m.Ttl != nil { + return *m.Ttl + } + return 0 +} + type JoinMessage struct { IsInit *bool `protobuf:"varint,1,opt,name=is_init" json:"is_init,omitempty"` Ip *string `protobuf:"bytes,2,req,name=ip" json:"ip,omitempty"` diff --git a/go/operation/system_message_test.go b/go/operation/system_message_test.go index 2731d0b2f..b5624c258 100644 --- a/go/operation/system_message_test.go +++ b/go/operation/system_message_test.go @@ -1,10 +1,11 @@ package operation import ( - proto "code.google.com/p/goprotobuf/proto" "encoding/json" "log" "testing" + + proto "code.google.com/p/goprotobuf/proto" ) func TestSerialDeserial(t *testing.T) { diff --git a/go/operation/upload_content.go b/go/operation/upload_content.go index b89e65ce8..480d76dca 100644 --- a/go/operation/upload_content.go +++ b/go/operation/upload_content.go @@ -2,7 +2,6 @@ package operation import ( "bytes" - "code.google.com/p/weed-fs/go/glog" "encoding/json" "errors" "fmt" @@ -14,6 +13,8 @@ import ( "net/textproto" "path/filepath" "strings" + + "github.com/chrislusf/weed-fs/go/glog" ) type UploadResult struct { diff --git a/go/proto/system_message.proto b/go/proto/system_message.proto index 15574ad56..ecd4973f7 100644 --- a/go/proto/system_message.proto +++ b/go/proto/system_message.proto @@ -10,6 +10,7 @@ message VolumeInformationMessage { optional bool read_only = 7; required uint32 replica_placement = 8; optional uint32 version = 9 [default=2]; + optional uint32 ttl = 10; } message JoinMessage { diff --git a/go/sequence/sequence.go b/go/sequence/sequence.go index 5a1bceaaf..1aa167b6b 100644 --- a/go/sequence/sequence.go +++ b/go/sequence/sequence.go @@ -1,7 +1,5 @@ package sequence -import () - type Sequencer interface { NextFileId(count int) (uint64, int) SetMax(uint64) diff --git a/go/stats/disk.go b/go/stats/disk.go index d5275e571..46d8c465e 100644 --- a/go/stats/disk.go +++ b/go/stats/disk.go @@ -1,7 +1,5 @@ package stats -import () - type DiskStatus struct { Dir string All uint64 diff --git a/go/stats/disk_notsupported.go b/go/stats/disk_notsupported.go index 37f9bc47d..e380d27ea 100644 --- a/go/stats/disk_notsupported.go +++ b/go/stats/disk_notsupported.go @@ -2,8 +2,6 @@ package stats -import () - func (disk *DiskStatus) fillInStatus() { return } diff --git a/go/stats/memory_notsupported.go b/go/stats/memory_notsupported.go index 64c3d7c2f..ba8229364 100644 --- a/go/stats/memory_notsupported.go +++ b/go/stats/memory_notsupported.go @@ -2,8 +2,6 @@ package stats -import () - func (mem *MemStatus) fillInStatus() { return } diff --git a/go/storage/cdb_map.go b/go/storage/cdb_map.go index d09a87e2a..fbb59e9c0 100644 --- a/go/storage/cdb_map.go +++ b/go/storage/cdb_map.go @@ -1,13 +1,14 @@ package storage import ( - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "fmt" - "github.com/tgulacsi/go-cdb" "os" "path/filepath" + + "github.com/chrislusf/weed-fs/go/util" + "github.com/tgulacsi/go-cdb" ) // CDB-backed read-only needle map diff --git a/go/storage/cdb_map_test.go b/go/storage/cdb_map_test.go index f6a7d42ad..ed690f44f 100644 --- a/go/storage/cdb_map_test.go +++ b/go/storage/cdb_map_test.go @@ -1,11 +1,12 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" "math/rand" "os" "runtime" "testing" + + "github.com/chrislusf/weed-fs/go/glog" ) var testIndexFilename string = "../../test/sample.idx" diff --git a/go/storage/compact_map.go b/go/storage/compact_map.go index 9cfc3e8f7..6ac18b012 100644 --- a/go/storage/compact_map.go +++ b/go/storage/compact_map.go @@ -1,7 +1,5 @@ package storage -import () - type NeedleValue struct { Key Key Offset uint32 `comment:"Volume offset"` //since aligned to 8 bytes, range is 4G*8=32G diff --git a/go/storage/compact_map_perf_test.go b/go/storage/compact_map_perf_test.go index 37b23a59f..f74684225 100644 --- a/go/storage/compact_map_perf_test.go +++ b/go/storage/compact_map_perf_test.go @@ -1,11 +1,12 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" "log" "os" "testing" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" ) func TestMemoryUsage(t *testing.T) { diff --git a/go/storage/compress.go b/go/storage/compress.go index 846fd0714..5efc9e1f1 100644 --- a/go/storage/compress.go +++ b/go/storage/compress.go @@ -2,11 +2,12 @@ package storage import ( "bytes" - "code.google.com/p/weed-fs/go/glog" "compress/flate" "compress/gzip" "io/ioutil" "strings" + + "github.com/chrislusf/weed-fs/go/glog" ) /* diff --git a/go/storage/crc.go b/go/storage/crc.go index 41f7f6d00..af25b9e53 100644 --- a/go/storage/crc.go +++ b/go/storage/crc.go @@ -1,9 +1,10 @@ package storage import ( - "code.google.com/p/weed-fs/go/util" "fmt" "hash/crc32" + + "github.com/chrislusf/weed-fs/go/util" ) var table = crc32.MakeTable(crc32.Castagnoli) @@ -25,5 +26,5 @@ func (c CRC) Value() uint32 { func (n *Needle) Etag() string { bits := make([]byte, 4) util.Uint32toBytes(bits, uint32(n.Checksum)) - return fmt.Sprintf("%x", bits) + return fmt.Sprintf("\"%x\"", bits) } diff --git a/go/storage/file_id.go b/go/storage/file_id.go index 5fcd8c387..f6e36a98c 100644 --- a/go/storage/file_id.go +++ b/go/storage/file_id.go @@ -1,11 +1,12 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" "encoding/hex" "errors" "strings" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" ) type FileId struct { diff --git a/go/storage/needle.go b/go/storage/needle.go index 77aa70169..aa3206920 100644 --- a/go/storage/needle.go +++ b/go/storage/needle.go @@ -1,9 +1,6 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/images" - "code.google.com/p/weed-fs/go/util" "encoding/hex" "errors" "io/ioutil" @@ -13,6 +10,10 @@ import ( "strconv" "strings" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/images" + "github.com/chrislusf/weed-fs/go/util" ) const ( @@ -38,12 +39,13 @@ type Needle struct { MimeSize uint8 //version2 Mime []byte `comment:"maximum 256 characters"` //version2 LastModified uint64 //only store LastModifiedBytesLength bytes, which is 5 bytes to disk + Ttl *TTL Checksum CRC `comment:"CRC32 to check integrity"` Padding []byte `comment:"Aligned to 8 bytes"` } -func ParseUpload(r *http.Request) (fileName string, data []byte, mimeType string, isGzipped bool, modifiedTime uint64, e error) { +func ParseUpload(r *http.Request) (fileName string, data []byte, mimeType string, isGzipped bool, modifiedTime uint64, ttl *TTL, e error) { form, fe := r.MultipartReader() if fe != nil { glog.V(0).Infoln("MultipartReader [ERROR]", fe) @@ -92,12 +94,13 @@ func ParseUpload(r *http.Request) (fileName string, data []byte, mimeType string fileName = fileName[:len(fileName)-3] } modifiedTime, _ = strconv.ParseUint(r.FormValue("ts"), 10, 64) + ttl, _ = ReadTTL(r.FormValue("ttl")) return } func NewNeedle(r *http.Request, fixJpgOrientation bool) (n *Needle, e error) { fname, mimeType, isGzipped := "", "", false n = new(Needle) - fname, n.Data, mimeType, isGzipped, n.LastModified, e = ParseUpload(r) + fname, n.Data, mimeType, isGzipped, n.LastModified, n.Ttl, e = ParseUpload(r) if e != nil { return } @@ -116,6 +119,9 @@ func NewNeedle(r *http.Request, fixJpgOrientation bool) (n *Needle, e error) { n.LastModified = uint64(time.Now().Unix()) } n.SetHasLastModifiedDate() + if n.Ttl != EMPTY_TTL { + n.SetHasTtl() + } if fixJpgOrientation { loweredName := strings.ToLower(fname) diff --git a/go/storage/needle_map.go b/go/storage/needle_map.go index 6d94ee1ca..504ca1552 100644 --- a/go/storage/needle_map.go +++ b/go/storage/needle_map.go @@ -1,11 +1,12 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" "fmt" "io" "os" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" ) type NeedleMapper interface { diff --git a/go/storage/needle_read_write.go b/go/storage/needle_read_write.go index 835d7c270..663b5abbd 100644 --- a/go/storage/needle_read_write.go +++ b/go/storage/needle_read_write.go @@ -1,12 +1,13 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" "errors" "fmt" "io" "os" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" ) const ( @@ -14,7 +15,9 @@ const ( FlagHasName = 0x02 FlagHasMime = 0x04 FlagHasLastModifiedDate = 0x08 + FlagHasTtl = 0x10 LastModifiedBytesLength = 5 + TtlBytesLength = 2 ) func (n *Needle) DiskSize() int64 { @@ -70,6 +73,9 @@ func (n *Needle) Append(w io.Writer, version Version) (size uint32, err error) { if n.HasLastModifiedDate() { n.Size = n.Size + LastModifiedBytesLength } + if n.HasTtl() { + n.Size = n.Size + TtlBytesLength + } } size = n.DataSize util.Uint32toBytes(header[12:16], n.Size) @@ -112,6 +118,12 @@ func (n *Needle) Append(w io.Writer, version Version) (size uint32, err error) { return } } + if n.HasTtl() { + n.Ttl.ToBytes(header[0:TtlBytesLength]) + if _, err = w.Write(header[0:TtlBytesLength]); err != nil { + return + } + } } padding := NeedlePaddingSize - ((NeedleHeaderSize + n.Size + NeedleChecksumSize) % NeedlePaddingSize) util.Uint32toBytes(header[0:NeedleChecksumSize], n.Checksum.Value()) @@ -194,6 +206,10 @@ func (n *Needle) readNeedleDataVersion2(bytes []byte) { n.LastModified = util.BytesToUint64(bytes[index : index+LastModifiedBytesLength]) index = index + LastModifiedBytesLength } + if index < lenBytes && n.HasTtl() { + n.Ttl = LoadTTLFromBytes(bytes[index : index+TtlBytesLength]) + index = index + TtlBytesLength + } } func ReadNeedleHeader(r *os.File, version Version, offset int64) (n *Needle, bodyLength uint32, err error) { @@ -263,3 +279,9 @@ func (n *Needle) HasLastModifiedDate() bool { func (n *Needle) SetHasLastModifiedDate() { n.Flags = n.Flags | FlagHasLastModifiedDate } +func (n *Needle) HasTtl() bool { + return n.Flags&FlagHasTtl > 0 +} +func (n *Needle) SetHasTtl() { + n.Flags = n.Flags | FlagHasTtl +} diff --git a/go/storage/replica_placement.go b/go/storage/replica_placement.go index 696888cd8..c1aca52eb 100644 --- a/go/storage/replica_placement.go +++ b/go/storage/replica_placement.go @@ -5,10 +5,6 @@ import ( "fmt" ) -const ( - ReplicaPlacementCount = 9 -) - type ReplicaPlacement struct { SameRackCount int DiffRackCount int @@ -55,7 +51,3 @@ func (rp *ReplicaPlacement) String() string { func (rp *ReplicaPlacement) GetCopyCount() int { return rp.DiffDataCenterCount + rp.DiffRackCount + rp.SameRackCount + 1 } - -func (rp *ReplicaPlacement) GetReplicationLevelIndex() int { - return rp.DiffDataCenterCount*3 + rp.DiffRackCount*3 + rp.SameRackCount -} diff --git a/go/storage/store.go b/go/storage/store.go index a6a4f399e..80d8a30b8 100644 --- a/go/storage/store.go +++ b/go/storage/store.go @@ -1,10 +1,6 @@ package storage import ( - proto "code.google.com/p/goprotobuf/proto" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "fmt" @@ -12,6 +8,15 @@ import ( "math/rand" "strconv" "strings" + + proto "code.google.com/p/goprotobuf/proto" + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/util" +) + +const ( + MAX_TTL_VOLUME_REMOVAL_DELAY = 10 // 10 minutes ) type DiskLocation struct { @@ -83,11 +88,15 @@ func NewStore(port int, ip, publicUrl string, dirnames []string, maxVolumeCounts } return } -func (s *Store) AddVolume(volumeListString string, collection string, replicaPlacement string) error { +func (s *Store) AddVolume(volumeListString string, collection string, replicaPlacement string, ttlString string) error { rt, e := NewReplicaPlacementFromString(replicaPlacement) if e != nil { return e } + ttl, e := ReadTTL(ttlString) + if e != nil { + return e + } for _, range_string := range strings.Split(volumeListString, ",") { if strings.Index(range_string, "-") < 0 { id_string := range_string @@ -95,7 +104,7 @@ func (s *Store) AddVolume(volumeListString string, collection string, replicaPla if err != nil { return fmt.Errorf("Volume Id %s is not a valid unsigned integer!", id_string) } - e = s.addVolume(VolumeId(id), collection, rt) + e = s.addVolume(VolumeId(id), collection, rt, ttl) } else { pair := strings.Split(range_string, "-") start, start_err := strconv.ParseUint(pair[0], 10, 64) @@ -107,7 +116,7 @@ func (s *Store) AddVolume(volumeListString string, collection string, replicaPla return fmt.Errorf("Volume End Id %s is not a valid unsigned integer!", pair[1]) } for id := start; id <= end; id++ { - if err := s.addVolume(VolumeId(id), collection, rt); err != nil { + if err := s.addVolume(VolumeId(id), collection, rt, ttl); err != nil { e = err } } @@ -129,6 +138,14 @@ func (s *Store) DeleteCollection(collection string) (e error) { } return } +func (s *Store) DeleteVolume(volumes map[VolumeId]*Volume, v *Volume) (e error) { + e = v.Destroy() + if e != nil { + return + } + delete(volumes, v.Id) + return +} func (s *Store) findVolume(vid VolumeId) *Volume { for _, location := range s.Locations { if v, found := location.volumes[vid]; found { @@ -148,13 +165,14 @@ func (s *Store) findFreeLocation() (ret *DiskLocation) { } return ret } -func (s *Store) addVolume(vid VolumeId, collection string, replicaPlacement *ReplicaPlacement) error { +func (s *Store) addVolume(vid VolumeId, collection string, replicaPlacement *ReplicaPlacement, ttl *TTL) error { if s.findVolume(vid) != nil { return fmt.Errorf("Volume Id %d already exists!", vid) } if location := s.findFreeLocation(); location != nil { - glog.V(0).Infoln("In dir", location.Directory, "adds volume =", vid, ", collection =", collection, ", replicaPlacement =", replicaPlacement) - if volume, err := NewVolume(location.Directory, collection, vid, replicaPlacement); err == nil { + glog.V(0).Infof("In dir %s adds volume:%v collection:%s replicaPlacement:%v ttl:%v", + location.Directory, vid, collection, replicaPlacement, ttl) + if volume, err := NewVolume(location.Directory, collection, vid, replicaPlacement, ttl); err == nil { location.volumes[vid] = volume return nil } else { @@ -190,9 +208,9 @@ func (l *DiskLocation) loadExistingVolumes() { } if vid, err := NewVolumeId(base); err == nil { if l.volumes[vid] == nil { - if v, e := NewVolume(l.Directory, collection, vid, nil); e == nil { + if v, e := NewVolume(l.Directory, collection, vid, nil, nil); e == nil { l.volumes[vid] = v - glog.V(0).Infoln("data file", l.Directory+"/"+name, "replicaPlacement =", v.ReplicaPlacement, "version =", v.Version(), "size =", v.Size()) + glog.V(0).Infof("data file %s, replicaPlacement=%s v=%d size=%d ttl=%s", l.Directory+"/"+name, v.ReplicaPlacement, v.Version(), v.Size(), v.Ttl.String()) } } } @@ -240,21 +258,31 @@ func (s *Store) Join() (masterNode string, e error) { for _, location := range s.Locations { maxVolumeCount = maxVolumeCount + location.MaxVolumeCount for k, v := range location.volumes { - volumeMessage := &operation.VolumeInformationMessage{ - Id: proto.Uint32(uint32(k)), - Size: proto.Uint64(uint64(v.Size())), - Collection: proto.String(v.Collection), - FileCount: proto.Uint64(uint64(v.nm.FileCount())), - DeleteCount: proto.Uint64(uint64(v.nm.DeletedCount())), - DeletedByteCount: proto.Uint64(v.nm.DeletedSize()), - ReadOnly: proto.Bool(v.readOnly), - ReplicaPlacement: proto.Uint32(uint32(v.ReplicaPlacement.Byte())), - Version: proto.Uint32(uint32(v.Version())), - } - volumeMessages = append(volumeMessages, volumeMessage) if maxFileKey < v.nm.MaxFileKey() { maxFileKey = v.nm.MaxFileKey() } + if !v.expired(s.volumeSizeLimit) { + volumeMessage := &operation.VolumeInformationMessage{ + Id: proto.Uint32(uint32(k)), + Size: proto.Uint64(uint64(v.Size())), + Collection: proto.String(v.Collection), + FileCount: proto.Uint64(uint64(v.nm.FileCount())), + DeleteCount: proto.Uint64(uint64(v.nm.DeletedCount())), + DeletedByteCount: proto.Uint64(v.nm.DeletedSize()), + ReadOnly: proto.Bool(v.readOnly), + ReplicaPlacement: proto.Uint32(uint32(v.ReplicaPlacement.Byte())), + Version: proto.Uint32(uint32(v.Version())), + Ttl: proto.Uint32(v.Ttl.ToUint32()), + } + volumeMessages = append(volumeMessages, volumeMessage) + } else { + if v.exiredLongEnough(MAX_TTL_VOLUME_REMOVAL_DELAY) { + s.DeleteVolume(location.volumes, v) + glog.V(0).Infoln("volume", v.Id, "is deleted.") + } else { + glog.V(0).Infoln("volume", v.Id, "is expired.") + } + } } } diff --git a/go/storage/store_vacuum.go b/go/storage/store_vacuum.go index 5adaa7561..209e3b4b3 100644 --- a/go/storage/store_vacuum.go +++ b/go/storage/store_vacuum.go @@ -1,9 +1,10 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" "fmt" "strconv" + + "github.com/chrislusf/weed-fs/go/glog" ) func (s *Store) CheckCompactVolume(volumeIdString string, garbageThresholdString string) (error, bool) { diff --git a/go/storage/volume.go b/go/storage/volume.go index 7bd8e7467..5b0a83605 100644 --- a/go/storage/volume.go +++ b/go/storage/volume.go @@ -2,7 +2,6 @@ package storage import ( "bytes" - "code.google.com/p/weed-fs/go/glog" "errors" "fmt" "io" @@ -10,24 +9,10 @@ import ( "path" "sync" "time" -) -const ( - SuperBlockSize = 8 + "github.com/chrislusf/weed-fs/go/glog" ) -type SuperBlock struct { - Version Version - ReplicaPlacement *ReplicaPlacement -} - -func (s *SuperBlock) Bytes() []byte { - header := make([]byte, SuperBlockSize) - header[0] = byte(s.Version) - header[1] = s.ReplicaPlacement.Byte() - return header -} - type Volume struct { Id VolumeId dir string @@ -38,12 +23,13 @@ type Volume struct { SuperBlock - accessLock sync.Mutex + accessLock sync.Mutex + lastModifiedTime uint64 //unix time in seconds } -func NewVolume(dirname string, collection string, id VolumeId, replicaPlacement *ReplicaPlacement) (v *Volume, e error) { +func NewVolume(dirname string, collection string, id VolumeId, replicaPlacement *ReplicaPlacement, ttl *TTL) (v *Volume, e error) { v = &Volume{dir: dirname, Collection: collection, Id: id} - v.SuperBlock = SuperBlock{ReplicaPlacement: replicaPlacement} + v.SuperBlock = SuperBlock{ReplicaPlacement: replicaPlacement, Ttl: ttl} e = v.load(true, true) return } @@ -65,12 +51,13 @@ func (v *Volume) load(alsoLoadIndex bool, createDatIfMissing bool) error { var e error fileName := v.FileName() - if exists, canRead, canWrite, _ := checkFile(fileName + ".dat"); exists { + if exists, canRead, canWrite, modifiedTime := checkFile(fileName + ".dat"); exists { if !canRead { return fmt.Errorf("cannot read Volume Data file %s.dat", fileName) } if canWrite { v.dataFile, e = os.OpenFile(fileName+".dat", os.O_RDWR|os.O_CREATE, 0644) + v.lastModifiedTime = uint64(modifiedTime.Unix()) } else { glog.V(0).Infoln("opening " + fileName + ".dat in READONLY mode") v.dataFile, e = os.Open(fileName + ".dat") @@ -86,7 +73,7 @@ func (v *Volume) load(alsoLoadIndex bool, createDatIfMissing bool) error { if e != nil { if !os.IsPermission(e) { - return fmt.Errorf("cannot load Volume Data %s.dat: %s", fileName, e.Error()) + return fmt.Errorf("cannot load Volume Data %s.dat: %v", fileName, e) } } @@ -106,12 +93,12 @@ func (v *Volume) load(alsoLoadIndex bool, createDatIfMissing bool) error { if v.readOnly { glog.V(1).Infoln("open to read file", fileName+".idx") if indexFile, e = os.OpenFile(fileName+".idx", os.O_RDONLY, 0644); e != nil { - return fmt.Errorf("cannot read Volume Index %s.idx: %s", fileName, e.Error()) + return fmt.Errorf("cannot read Volume Index %s.idx: %v", fileName, e) } } else { glog.V(1).Infoln("open to write file", fileName+".idx") if indexFile, e = os.OpenFile(fileName+".idx", os.O_RDWR|os.O_CREATE, 0644); e != nil { - return fmt.Errorf("cannot write Volume Index %s.idx: %s", fileName, e.Error()) + return fmt.Errorf("cannot write Volume Index %s.idx: %v", fileName, e) } } glog.V(0).Infoln("loading file", fileName+".idx", "readonly", v.readOnly) @@ -122,14 +109,14 @@ func (v *Volume) load(alsoLoadIndex bool, createDatIfMissing bool) error { return e } func (v *Volume) Version() Version { - return v.SuperBlock.Version + return v.SuperBlock.Version() } func (v *Volume) Size() int64 { stat, e := v.dataFile.Stat() if e == nil { return stat.Size() } - glog.V(0).Infof("Failed to read file size %s %s", v.dataFile.Name(), e.Error()) + glog.V(0).Infof("Failed to read file size %s %v", v.dataFile.Name(), e) return -1 } func (v *Volume) Close() { @@ -138,44 +125,6 @@ func (v *Volume) Close() { v.nm.Close() _ = v.dataFile.Close() } -func (v *Volume) maybeWriteSuperBlock() error { - stat, e := v.dataFile.Stat() - if e != nil { - glog.V(0).Infof("failed to stat datafile %s: %s", v.dataFile, e.Error()) - return e - } - if stat.Size() == 0 { - v.SuperBlock.Version = CurrentVersion - _, e = v.dataFile.Write(v.SuperBlock.Bytes()) - if e != nil && os.IsPermission(e) { - //read-only, but zero length - recreate it! - if v.dataFile, e = os.Create(v.dataFile.Name()); e == nil { - if _, e = v.dataFile.Write(v.SuperBlock.Bytes()); e == nil { - v.readOnly = false - } - } - } - } - return e -} -func (v *Volume) readSuperBlock() (err error) { - if _, err = v.dataFile.Seek(0, 0); err != nil { - return fmt.Errorf("cannot seek to the beginning of %s: %s", v.dataFile.Name(), err.Error()) - } - header := make([]byte, SuperBlockSize) - if _, e := v.dataFile.Read(header); e != nil { - return fmt.Errorf("cannot read superblock: %s", e.Error()) - } - v.SuperBlock, err = ParseSuperBlock(header) - return err -} -func ParseSuperBlock(header []byte) (superBlock SuperBlock, err error) { - superBlock.Version = Version(header[0]) - if superBlock.ReplicaPlacement, err = NewReplicaPlacementFromByte(header[1]); err != nil { - err = fmt.Errorf("cannot read replica type: %s", err.Error()) - } - return -} func (v *Volume) NeedToReplicate() bool { return v.ReplicaPlacement.GetCopyCount() > 1 } @@ -222,6 +171,7 @@ func (v *Volume) write(n *Needle) (size uint32, err error) { } var offset int64 if offset, err = v.dataFile.Seek(0, 2); err != nil { + glog.V(0).Infof("faile to seek the end of file: %v", err) return } @@ -229,23 +179,26 @@ func (v *Volume) write(n *Needle) (size uint32, err error) { if offset%NeedlePaddingSize != 0 { offset = offset + (NeedlePaddingSize - offset%NeedlePaddingSize) if offset, err = v.dataFile.Seek(offset, 0); err != nil { - glog.V(4).Infof("failed to align in datafile %s: %s", v.dataFile.Name(), err.Error()) + glog.V(0).Infof("failed to align in datafile %s: %v", v.dataFile.Name(), err) return } } if size, err = n.Append(v.dataFile, v.Version()); err != nil { if e := v.dataFile.Truncate(offset); e != nil { - err = fmt.Errorf("%s\ncannot truncate %s: %s", err, v.dataFile.Name(), e.Error()) + err = fmt.Errorf("%s\ncannot truncate %s: %v", err, v.dataFile.Name(), e) } return } nv, ok := v.nm.Get(n.Id) if !ok || int64(nv.Offset)*NeedlePaddingSize < offset { if _, err = v.nm.Put(n.Id, uint32(offset/NeedlePaddingSize), n.Size); err != nil { - glog.V(4).Infof("failed to save in needle map %d: %s", n.Id, err.Error()) + glog.V(4).Infof("failed to save in needle map %d: %v", n.Id, err) } } + if v.lastModifiedTime < n.LastModified { + v.lastModifiedTime = n.LastModified + } return } @@ -275,8 +228,25 @@ func (v *Volume) delete(n *Needle) (uint32, error) { func (v *Volume) read(n *Needle) (int, error) { nv, ok := v.nm.Get(n.Id) - if ok && nv.Offset > 0 { - return n.Read(v.dataFile, int64(nv.Offset)*NeedlePaddingSize, nv.Size, v.Version()) + if !ok || nv.Offset == 0 { + return -1, errors.New("Not Found") + } + bytesRead, err := n.Read(v.dataFile, int64(nv.Offset)*NeedlePaddingSize, nv.Size, v.Version()) + if err != nil { + return bytesRead, err + } + if !n.HasTtl() { + return bytesRead, err + } + ttlMinutes := n.Ttl.Minutes() + if ttlMinutes == 0 { + return bytesRead, nil + } + if !n.HasLastModifiedDate() { + return bytesRead, nil + } + if uint64(time.Now().Unix()) < n.LastModified+uint64(ttlMinutes*60) { + return bytesRead, nil } return -1, errors.New("Not Found") } @@ -324,13 +294,13 @@ func ScanVolumeFile(dirname string, collection string, id VolumeId, offset := int64(SuperBlockSize) n, rest, e := ReadNeedleHeader(v.dataFile, version, offset) if e != nil { - err = fmt.Errorf("cannot read needle header: %s", e) + err = fmt.Errorf("cannot read needle header: %v", e) return } for n != nil { if readNeedleBody { if err = n.ReadNeedleBody(v.dataFile, version, offset+int64(NeedleHeaderSize), rest); err != nil { - err = fmt.Errorf("cannot read needle body: %s", err) + err = fmt.Errorf("cannot read needle body: %v", err) return } } @@ -342,7 +312,7 @@ func ScanVolumeFile(dirname string, collection string, id VolumeId, if err == io.EOF { return nil } - return fmt.Errorf("cannot read needle header: %s", err) + return fmt.Errorf("cannot read needle header: %v", err) } } @@ -392,8 +362,48 @@ func (v *Volume) ensureConvertIdxToCdb(fileName string) (cdbCanRead bool) { defer indexFile.Close() glog.V(0).Infof("converting %s.idx to %s.cdb", fileName, fileName) if e = ConvertIndexToCdb(fileName+".cdb", indexFile); e != nil { - glog.V(0).Infof("error converting %s.idx to %s.cdb: %s", fileName, fileName, e.Error()) + glog.V(0).Infof("error converting %s.idx to %s.cdb: %v", fileName, fileName, e) return false } return true } + +// volume is expired if modified time + volume ttl < now +// except when volume is empty +// or when the volume does not have a ttl +// or when volumeSizeLimit is 0 when server just starts +func (v *Volume) expired(volumeSizeLimit uint64) bool { + if volumeSizeLimit == 0 { + //skip if we don't know size limit + return false + } + if v.ContentSize() == 0 { + return false + } + if v.Ttl == nil || v.Ttl.Minutes() == 0 { + return false + } + glog.V(0).Infof("now:%v lastModified:%v", time.Now().Unix(), v.lastModifiedTime) + livedMinutes := (time.Now().Unix() - int64(v.lastModifiedTime)) / 60 + glog.V(0).Infof("ttl:%v lived:%v", v.Ttl, livedMinutes) + if int64(v.Ttl.Minutes()) < livedMinutes { + return true + } + return false +} + +// wait either maxDelayMinutes or 10% of ttl minutes +func (v *Volume) exiredLongEnough(maxDelayMinutes uint32) bool { + if v.Ttl == nil || v.Ttl.Minutes() == 0 { + return false + } + removalDelay := v.Ttl.Minutes() / 10 + if removalDelay > maxDelayMinutes { + removalDelay = maxDelayMinutes + } + + if uint64(v.Ttl.Minutes()+removalDelay)*60+v.lastModifiedTime < uint64(time.Now().Unix()) { + return true + } + return false +} diff --git a/go/storage/volume_info.go b/go/storage/volume_info.go index 165af1a19..6410c1784 100644 --- a/go/storage/volume_info.go +++ b/go/storage/volume_info.go @@ -1,13 +1,14 @@ package storage import ( - "code.google.com/p/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/operation" ) type VolumeInfo struct { Id VolumeId Size uint64 ReplicaPlacement *ReplicaPlacement + Ttl *TTL Collection string Version Version FileCount int @@ -32,5 +33,6 @@ func NewVolumeInfo(m *operation.VolumeInformationMessage) (vi VolumeInfo, err er return vi, e } vi.ReplicaPlacement = rp + vi.Ttl = LoadTTLFromUint32(*m.Ttl) return vi, nil } diff --git a/go/storage/volume_super_block.go b/go/storage/volume_super_block.go new file mode 100644 index 000000000..57e0deea9 --- /dev/null +++ b/go/storage/volume_super_block.go @@ -0,0 +1,76 @@ +package storage + +import ( + "fmt" + "os" + + "github.com/chrislusf/weed-fs/go/glog" +) + +const ( + SuperBlockSize = 8 +) + +/* +* Super block currently has 8 bytes allocated for each volume. +* Byte 0: version, 1 or 2 +* Byte 1: Replica Placement strategy, 000, 001, 002, 010, etc +* Byte 2 and byte 3: Time to live. See TTL for definition +* Rest bytes: Reserved + */ +type SuperBlock struct { + version Version + ReplicaPlacement *ReplicaPlacement + Ttl *TTL +} + +func (s *SuperBlock) Version() Version { + return s.version +} +func (s *SuperBlock) Bytes() []byte { + header := make([]byte, SuperBlockSize) + header[0] = byte(s.version) + header[1] = s.ReplicaPlacement.Byte() + s.Ttl.ToBytes(header[2:4]) + return header +} + +func (v *Volume) maybeWriteSuperBlock() error { + stat, e := v.dataFile.Stat() + if e != nil { + glog.V(0).Infof("failed to stat datafile %s: %s", v.dataFile, e.Error()) + return e + } + if stat.Size() == 0 { + v.SuperBlock.version = CurrentVersion + _, e = v.dataFile.Write(v.SuperBlock.Bytes()) + if e != nil && os.IsPermission(e) { + //read-only, but zero length - recreate it! + if v.dataFile, e = os.Create(v.dataFile.Name()); e == nil { + if _, e = v.dataFile.Write(v.SuperBlock.Bytes()); e == nil { + v.readOnly = false + } + } + } + } + return e +} +func (v *Volume) readSuperBlock() (err error) { + if _, err = v.dataFile.Seek(0, 0); err != nil { + return fmt.Errorf("cannot seek to the beginning of %s: %s", v.dataFile.Name(), err.Error()) + } + header := make([]byte, SuperBlockSize) + if _, e := v.dataFile.Read(header); e != nil { + return fmt.Errorf("cannot read superblock: %s", e.Error()) + } + v.SuperBlock, err = ParseSuperBlock(header) + return err +} +func ParseSuperBlock(header []byte) (superBlock SuperBlock, err error) { + superBlock.version = Version(header[0]) + if superBlock.ReplicaPlacement, err = NewReplicaPlacementFromByte(header[1]); err != nil { + err = fmt.Errorf("cannot read replica type: %s", err.Error()) + } + superBlock.Ttl = LoadTTLFromBytes(header[2:4]) + return +} diff --git a/go/storage/volume_super_block_test.go b/go/storage/volume_super_block_test.go new file mode 100644 index 000000000..13db4b194 --- /dev/null +++ b/go/storage/volume_super_block_test.go @@ -0,0 +1,23 @@ +package storage + +import ( + "testing" +) + +func TestSuperBlockReadWrite(t *testing.T) { + rp, _ := NewReplicaPlacementFromByte(byte(001)) + ttl, _ := ReadTTL("15d") + s := &SuperBlock{ + version: CurrentVersion, + ReplicaPlacement: rp, + Ttl: ttl, + } + + bytes := s.Bytes() + + if !(bytes[2] == 15 && bytes[3] == Day) { + println("byte[2]:", bytes[2], "byte[3]:", bytes[3]) + t.Fail() + } + +} diff --git a/go/storage/volume_ttl.go b/go/storage/volume_ttl.go new file mode 100644 index 000000000..459ee55ba --- /dev/null +++ b/go/storage/volume_ttl.go @@ -0,0 +1,135 @@ +package storage + +import ( + "strconv" +) + +const ( + //stored unit types + Empty byte = iota + Minute + Hour + Day + Week + Month + Year +) + +type TTL struct { + count byte + unit byte +} + +var EMPTY_TTL = &TTL{} + +// translate a readable ttl to internal ttl +// Supports format example: +// 3m: 3 minutes +// 4h: 4 hours +// 5d: 5 days +// 6w: 6 weeks +// 7M: 7 months +// 8y: 8 years +func ReadTTL(ttlString string) (*TTL, error) { + if ttlString == "" { + return EMPTY_TTL, nil + } + ttlBytes := []byte(ttlString) + unitByte := ttlBytes[len(ttlBytes)-1] + countBytes := ttlBytes[0 : len(ttlBytes)-1] + if '0' <= unitByte && unitByte <= '9' { + countBytes = ttlBytes + unitByte = 'm' + } + count, err := strconv.Atoi(string(countBytes)) + unit := toStoredByte(unitByte) + return &TTL{count: byte(count), unit: unit}, err +} + +// read stored bytes to a ttl +func LoadTTLFromBytes(input []byte) (t *TTL) { + return &TTL{count: input[0], unit: input[1]} +} + +// read stored bytes to a ttl +func LoadTTLFromUint32(ttl uint32) (t *TTL) { + input := make([]byte, 2) + input[1] = byte(ttl) + input[0] = byte(ttl >> 8) + return LoadTTLFromBytes(input) +} + +// save stored bytes to an output with 2 bytes +func (t TTL) ToBytes(output []byte) { + output[0] = t.count + output[1] = t.unit +} + +func (t TTL) ToUint32() (output uint32) { + output = uint32(t.count) << 8 + output += uint32(t.unit) + return output +} + +func (t TTL) String() string { + if t.count == 0 { + return "" + } + if t.unit == Empty { + return "" + } + countString := strconv.Itoa(int(t.count)) + switch t.unit { + case Minute: + return countString + "m" + case Hour: + return countString + "h" + case Day: + return countString + "d" + case Week: + return countString + "w" + case Month: + return countString + "M" + case Year: + return countString + "y" + } + return "" +} + +func toStoredByte(readableUnitByte byte) byte { + switch readableUnitByte { + case 'm': + return Minute + case 'h': + return Hour + case 'd': + return Day + case 'w': + return Week + case 'M': + return Month + case 'y': + return Year + } + return 0 +} + +func (t TTL) Minutes() uint32 { + switch t.unit { + case Empty: + return 0 + case Minute: + return uint32(t.count) + case Hour: + return uint32(t.count) * 60 + case Day: + return uint32(t.count) * 60 * 24 + case Week: + return uint32(t.count) * 60 * 24 * 7 + case Month: + return uint32(t.count) * 60 * 24 * 31 + case Year: + return uint32(t.count) * 60 * 24 * 365 + } + return 0 +} diff --git a/go/storage/volume_ttl_test.go b/go/storage/volume_ttl_test.go new file mode 100644 index 000000000..216469a4c --- /dev/null +++ b/go/storage/volume_ttl_test.go @@ -0,0 +1,60 @@ +package storage + +import ( + "testing" +) + +func TestTTLReadWrite(t *testing.T) { + ttl, _ := ReadTTL("") + if ttl.Minutes() != 0 { + t.Errorf("empty ttl:%v", ttl) + } + + ttl, _ = ReadTTL("9") + if ttl.Minutes() != 9 { + t.Errorf("9 ttl:%v", ttl) + } + + ttl, _ = ReadTTL("8m") + if ttl.Minutes() != 8 { + t.Errorf("8m ttl:%v", ttl) + } + + ttl, _ = ReadTTL("5h") + if ttl.Minutes() != 300 { + t.Errorf("5h ttl:%v", ttl) + } + + ttl, _ = ReadTTL("5d") + if ttl.Minutes() != 5*24*60 { + t.Errorf("5d ttl:%v", ttl) + } + + ttl, _ = ReadTTL("5w") + if ttl.Minutes() != 5*7*24*60 { + t.Errorf("5w ttl:%v", ttl) + } + + ttl, _ = ReadTTL("5M") + if ttl.Minutes() != 5*31*24*60 { + t.Errorf("5M ttl:%v", ttl) + } + + ttl, _ = ReadTTL("5y") + if ttl.Minutes() != 5*365*24*60 { + t.Errorf("5y ttl:%v", ttl) + } + + output := make([]byte, 2) + ttl.ToBytes(output) + ttl2 := LoadTTLFromBytes(output) + if ttl.Minutes() != ttl2.Minutes() { + t.Errorf("ttl:%v ttl2:%v", ttl, ttl2) + } + + ttl3 := LoadTTLFromUint32(ttl.ToUint32()) + if ttl.Minutes() != ttl3.Minutes() { + t.Errorf("ttl:%v ttl3:%v", ttl, ttl3) + } + +} diff --git a/go/storage/volume_vacuum.go b/go/storage/volume_vacuum.go index 7d2a38cb8..7e026a61d 100644 --- a/go/storage/volume_vacuum.go +++ b/go/storage/volume_vacuum.go @@ -1,10 +1,11 @@ package storage import ( - "code.google.com/p/weed-fs/go/glog" "fmt" "os" - _ "time" + "time" + + "github.com/chrislusf/weed-fs/go/glog" ) func (v *Volume) garbageLevel() float64 { @@ -13,9 +14,10 @@ func (v *Volume) garbageLevel() float64 { func (v *Volume) Compact() error { glog.V(3).Infof("Compacting ...") - v.accessLock.Lock() - defer v.accessLock.Unlock() - glog.V(3).Infof("Got Compaction lock...") + //no need to lock for copy on write + //v.accessLock.Lock() + //defer v.accessLock.Unlock() + //glog.V(3).Infof("Got Compaction lock...") filePath := v.FileName() glog.V(3).Infof("creating copies for volume %d ...", v.Id) @@ -59,10 +61,15 @@ func (v *Volume) copyDataAndGenerateIndexFile(dstName, idxName string) (err erro nm := NewNeedleMap(idx) new_offset := int64(SuperBlockSize) + now := uint64(time.Now().Unix()) + err = ScanVolumeFile(v.dir, v.Collection, v.Id, func(superBlock SuperBlock) error { _, err = dst.Write(superBlock.Bytes()) return err }, true, func(n *Needle, offset int64) error { + if n.HasTtl() && now >= n.LastModified+uint64(v.Ttl.Minutes()*60) { + return nil + } nv, ok := v.nm.Get(n.Id) glog.V(4).Infoln("needle expected offset ", offset, "ok", ok, "nv", nv) if ok && int64(nv.Offset)*NeedlePaddingSize == offset && nv.Size > 0 { diff --git a/go/storage/volume_version.go b/go/storage/volume_version.go index 9702ae904..2e9f58aa2 100644 --- a/go/storage/volume_version.go +++ b/go/storage/volume_version.go @@ -1,7 +1,5 @@ package storage -import () - type Version uint8 const ( diff --git a/go/tools/read_index.go b/go/tools/read_index.go index a958de410..1104dc348 100644 --- a/go/tools/read_index.go +++ b/go/tools/read_index.go @@ -1,11 +1,12 @@ package main import ( - "code.google.com/p/weed-fs/go/storage" "flag" "fmt" "log" "os" + + "github.com/chrislusf/weed-fs/go/storage" ) var ( diff --git a/go/topology/allocate_volume.go b/go/topology/allocate_volume.go index 77b4ac508..a791b4c1c 100644 --- a/go/topology/allocate_volume.go +++ b/go/topology/allocate_volume.go @@ -1,22 +1,24 @@ package topology import ( - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "net/url" + + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" ) type AllocateVolumeResult struct { Error string } -func AllocateVolume(dn *DataNode, vid storage.VolumeId, collection string, rp *storage.ReplicaPlacement) error { +func AllocateVolume(dn *DataNode, vid storage.VolumeId, option *VolumeGrowOption) error { values := make(url.Values) values.Add("volume", vid.String()) - values.Add("collection", collection) - values.Add("replication", rp.String()) + values.Add("collection", option.Collection) + values.Add("replication", option.ReplicaPlacement.String()) + values.Add("ttl", option.Ttl.String()) jsonBlob, err := util.Post("http://"+dn.PublicUrl+"/admin/assign_volume", values) if err != nil { return err diff --git a/go/topology/cluster_commands.go b/go/topology/cluster_commands.go index 703435173..cafc52c76 100644 --- a/go/topology/cluster_commands.go +++ b/go/topology/cluster_commands.go @@ -1,8 +1,8 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" "github.com/goraft/raft" ) diff --git a/go/topology/collection.go b/go/topology/collection.go index b21122d22..4b47ae88a 100644 --- a/go/topology/collection.go +++ b/go/topology/collection.go @@ -1,35 +1,37 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" ) type Collection struct { Name string volumeSizeLimit uint64 - replicaType2VolumeLayout []*VolumeLayout + storageType2VolumeLayout *util.ConcurrentReadMap } func NewCollection(name string, volumeSizeLimit uint64) *Collection { c := &Collection{Name: name, volumeSizeLimit: volumeSizeLimit} - c.replicaType2VolumeLayout = make([]*VolumeLayout, storage.ReplicaPlacementCount) + c.storageType2VolumeLayout = util.NewConcurrentReadMap() return c } -func (c *Collection) GetOrCreateVolumeLayout(rp *storage.ReplicaPlacement) *VolumeLayout { - replicaPlacementIndex := rp.GetReplicationLevelIndex() - if c.replicaType2VolumeLayout[replicaPlacementIndex] == nil { - glog.V(0).Infoln("collection", c.Name, "adding replication type", rp) - c.replicaType2VolumeLayout[replicaPlacementIndex] = NewVolumeLayout(rp, c.volumeSizeLimit) +func (c *Collection) GetOrCreateVolumeLayout(rp *storage.ReplicaPlacement, ttl *storage.TTL) *VolumeLayout { + keyString := rp.String() + if ttl != nil { + keyString += ttl.String() } - return c.replicaType2VolumeLayout[replicaPlacementIndex] + vl := c.storageType2VolumeLayout.Get(keyString, func() interface{} { + return NewVolumeLayout(rp, ttl, c.volumeSizeLimit) + }) + return vl.(*VolumeLayout) } func (c *Collection) Lookup(vid storage.VolumeId) []*DataNode { - for _, vl := range c.replicaType2VolumeLayout { + for _, vl := range c.storageType2VolumeLayout.Items { if vl != nil { - if list := vl.Lookup(vid); list != nil { + if list := vl.(*VolumeLayout).Lookup(vid); list != nil { return list } } @@ -38,9 +40,9 @@ func (c *Collection) Lookup(vid storage.VolumeId) []*DataNode { } func (c *Collection) ListVolumeServers() (nodes []*DataNode) { - for _, vl := range c.replicaType2VolumeLayout { + for _, vl := range c.storageType2VolumeLayout.Items { if vl != nil { - if list := vl.ListVolumeServers(); list != nil { + if list := vl.(*VolumeLayout).ListVolumeServers(); list != nil { nodes = append(nodes, list...) } } diff --git a/go/topology/data_center.go b/go/topology/data_center.go index ebd07803b..bcf2dfd31 100644 --- a/go/topology/data_center.go +++ b/go/topology/data_center.go @@ -1,7 +1,5 @@ package topology -import () - type DataCenter struct { NodeImpl } diff --git a/go/topology/data_node.go b/go/topology/data_node.go index ae80e08bb..109bd037f 100644 --- a/go/topology/data_node.go +++ b/go/topology/data_node.go @@ -1,9 +1,10 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "strconv" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) type DataNode struct { @@ -38,15 +39,16 @@ func (dn *DataNode) AddOrUpdateVolume(v storage.VolumeInfo) { } } -func (dn *DataNode) UpdateVolumes(actualVolumes []storage.VolumeInfo) { +func (dn *DataNode) UpdateVolumes(actualVolumes []storage.VolumeInfo) (deletedVolumes []storage.VolumeInfo) { actualVolumeMap := make(map[storage.VolumeId]storage.VolumeInfo) for _, v := range actualVolumes { actualVolumeMap[v.Id] = v } - for vid, _ := range dn.volumes { + for vid, v := range dn.volumes { if _, ok := actualVolumeMap[vid]; !ok { glog.V(0).Infoln("Deleting volume id:", vid) delete(dn.volumes, vid) + deletedVolumes = append(deletedVolumes, v) dn.UpAdjustVolumeCountDelta(-1) dn.UpAdjustActiveVolumeCountDelta(-1) } @@ -54,6 +56,7 @@ func (dn *DataNode) UpdateVolumes(actualVolumes []storage.VolumeInfo) { for _, v := range actualVolumes { dn.AddOrUpdateVolume(v) } + return } func (dn *DataNode) GetDataCenter() *DataCenter { diff --git a/go/topology/node.go b/go/topology/node.go index c52414008..10955fa72 100644 --- a/go/topology/node.go +++ b/go/topology/node.go @@ -1,11 +1,12 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "errors" "math/rand" "strings" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) type NodeId string diff --git a/go/topology/store_replicate.go b/go/topology/store_replicate.go index a982cebe5..0c52f9d30 100644 --- a/go/topology/store_replicate.go +++ b/go/topology/store_replicate.go @@ -2,12 +2,13 @@ package topology import ( "bytes" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/util" "net/http" "strconv" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" ) func ReplicatedWrite(masterNode string, s *storage.Store, volumeId storage.VolumeId, needle *storage.Needle, r *http.Request) (size uint32, errorStatus string) { diff --git a/go/topology/topology.go b/go/topology/topology.go index f1daffb53..c2073ed2f 100644 --- a/go/topology/topology.go +++ b/go/topology/topology.go @@ -1,20 +1,22 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/sequence" - "code.google.com/p/weed-fs/go/storage" "errors" - "github.com/goraft/raft" "io/ioutil" "math/rand" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/sequence" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" + "github.com/goraft/raft" ) type Topology struct { NodeImpl - collectionMap map[string]*Collection + collectionMap *util.ConcurrentReadMap pulse int64 @@ -37,7 +39,7 @@ func NewTopology(id string, confFile string, seq sequence.Sequencer, volumeSizeL t.nodeType = "Topology" t.NodeImpl.value = t t.children = make(map[NodeId]Node) - t.collectionMap = make(map[string]*Collection) + t.collectionMap = util.NewConcurrentReadMap() t.pulse = int64(pulse) t.volumeSizeLimit = volumeSizeLimit @@ -89,14 +91,14 @@ func (t *Topology) loadConfiguration(configurationFile string) error { func (t *Topology) Lookup(collection string, vid storage.VolumeId) []*DataNode { //maybe an issue if lots of collections? if collection == "" { - for _, c := range t.collectionMap { - if list := c.Lookup(vid); list != nil { + for _, c := range t.collectionMap.Items { + if list := c.(*Collection).Lookup(vid); list != nil { return list } } } else { - if c, ok := t.collectionMap[collection]; ok { - return c.Lookup(vid) + if c, ok := t.collectionMap.Items[collection]; ok { + return c.(*Collection).Lookup(vid) } } return nil @@ -109,13 +111,13 @@ func (t *Topology) NextVolumeId() storage.VolumeId { return next } -func (t *Topology) HasWriableVolume(option *VolumeGrowOption) bool { - vl := t.GetVolumeLayout(option.Collection, option.ReplicaPlacement) +func (t *Topology) HasWritableVolume(option *VolumeGrowOption) bool { + vl := t.GetVolumeLayout(option.Collection, option.ReplicaPlacement, option.Ttl) return vl.GetActiveVolumeCount(option) > 0 } func (t *Topology) PickForWrite(count int, option *VolumeGrowOption) (string, int, *DataNode, error) { - vid, count, datanodes, err := t.GetVolumeLayout(option.Collection, option.ReplicaPlacement).PickForWrite(count, option) + vid, count, datanodes, err := t.GetVolumeLayout(option.Collection, option.ReplicaPlacement, option.Ttl).PickForWrite(count, option) if err != nil || datanodes.Length() == 0 { return "", 0, nil, errors.New("No writable volumes avalable!") } @@ -123,28 +125,30 @@ func (t *Topology) PickForWrite(count int, option *VolumeGrowOption) (string, in return storage.NewFileId(*vid, fileId, rand.Uint32()).String(), count, datanodes.Head(), nil } -func (t *Topology) GetVolumeLayout(collectionName string, rp *storage.ReplicaPlacement) *VolumeLayout { - _, ok := t.collectionMap[collectionName] - if !ok { - t.collectionMap[collectionName] = NewCollection(collectionName, t.volumeSizeLimit) - } - return t.collectionMap[collectionName].GetOrCreateVolumeLayout(rp) +func (t *Topology) GetVolumeLayout(collectionName string, rp *storage.ReplicaPlacement, ttl *storage.TTL) *VolumeLayout { + return t.collectionMap.Get(collectionName, func() interface{} { + return NewCollection(collectionName, t.volumeSizeLimit) + }).(*Collection).GetOrCreateVolumeLayout(rp, ttl) } -func (t *Topology) GetCollection(collectionName string) (collection *Collection, ok bool) { - collection, ok = t.collectionMap[collectionName] - return +func (t *Topology) GetCollection(collectionName string) (*Collection, bool) { + c, hasCollection := t.collectionMap.Items[collectionName] + return c.(*Collection), hasCollection } func (t *Topology) DeleteCollection(collectionName string) { - delete(t.collectionMap, collectionName) + delete(t.collectionMap.Items, collectionName) } func (t *Topology) RegisterVolumeLayout(v storage.VolumeInfo, dn *DataNode) { - t.GetVolumeLayout(v.Collection, v.ReplicaPlacement).RegisterVolume(&v, dn) + t.GetVolumeLayout(v.Collection, v.ReplicaPlacement, v.Ttl).RegisterVolume(&v, dn) +} +func (t *Topology) UnRegisterVolumeLayout(v storage.VolumeInfo, dn *DataNode) { + glog.Infof("removing volume info:%+v", v) + t.GetVolumeLayout(v.Collection, v.ReplicaPlacement, v.Ttl).UnRegisterVolume(&v, dn) } -func (t *Topology) RegisterVolumes(joinMessage *operation.JoinMessage) { +func (t *Topology) ProcessJoinMessage(joinMessage *operation.JoinMessage) { t.Sequence.SetMax(*joinMessage.MaxFileKey) dcName, rackName := t.configuration.Locate(*joinMessage.Ip, *joinMessage.DataCenter, *joinMessage.Rack) dc := t.GetOrCreateDataCenter(dcName) @@ -162,10 +166,13 @@ func (t *Topology) RegisterVolumes(joinMessage *operation.JoinMessage) { glog.V(0).Infoln("Fail to convert joined volume information:", err.Error()) } } - dn.UpdateVolumes(volumeInfos) + deletedVolumes := dn.UpdateVolumes(volumeInfos) for _, v := range volumeInfos { t.RegisterVolumeLayout(v, dn) } + for _, v := range deletedVolumes { + t.UnRegisterVolumeLayout(v, dn) + } } func (t *Topology) GetOrCreateDataCenter(dcName string) *DataCenter { diff --git a/go/topology/topology_event_handling.go b/go/topology/topology_event_handling.go index 7398ff9bf..7e36568b6 100644 --- a/go/topology/topology_event_handling.go +++ b/go/topology/topology_event_handling.go @@ -1,10 +1,11 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "math/rand" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) func (t *Topology) StartRefreshWritableVolumes(garbageThreshold string) { @@ -41,7 +42,7 @@ func (t *Topology) StartRefreshWritableVolumes(garbageThreshold string) { }() } func (t *Topology) SetVolumeCapacityFull(volumeInfo storage.VolumeInfo) bool { - vl := t.GetVolumeLayout(volumeInfo.Collection, volumeInfo.ReplicaPlacement) + vl := t.GetVolumeLayout(volumeInfo.Collection, volumeInfo.ReplicaPlacement, volumeInfo.Ttl) if !vl.SetVolumeCapacityFull(volumeInfo.Id) { return false } @@ -55,7 +56,7 @@ func (t *Topology) SetVolumeCapacityFull(volumeInfo storage.VolumeInfo) bool { func (t *Topology) UnRegisterDataNode(dn *DataNode) { for _, v := range dn.volumes { glog.V(0).Infoln("Removing Volume", v.Id, "from the dead volume server", dn) - vl := t.GetVolumeLayout(v.Collection, v.ReplicaPlacement) + vl := t.GetVolumeLayout(v.Collection, v.ReplicaPlacement, v.Ttl) vl.SetVolumeUnavailable(dn, v.Id) } dn.UpAdjustVolumeCountDelta(-dn.GetVolumeCount()) @@ -65,7 +66,7 @@ func (t *Topology) UnRegisterDataNode(dn *DataNode) { } func (t *Topology) RegisterRecoveredDataNode(dn *DataNode) { for _, v := range dn.volumes { - vl := t.GetVolumeLayout(v.Collection, v.ReplicaPlacement) + vl := t.GetVolumeLayout(v.Collection, v.ReplicaPlacement, v.Ttl) if vl.isWritable(&v) { vl.SetVolumeAvailable(dn, v.Id) } diff --git a/go/topology/topology_map.go b/go/topology/topology_map.go index f66d4c251..6a1423ca8 100644 --- a/go/topology/topology_map.go +++ b/go/topology/topology_map.go @@ -1,7 +1,5 @@ package topology -import () - func (t *Topology) ToMap() interface{} { m := make(map[string]interface{}) m["Max"] = t.GetMaxVolumeCount() @@ -13,10 +11,11 @@ func (t *Topology) ToMap() interface{} { } m["DataCenters"] = dcs var layouts []interface{} - for _, c := range t.collectionMap { - for _, layout := range c.replicaType2VolumeLayout { + for _, col := range t.collectionMap.Items { + c := col.(*Collection) + for _, layout := range c.storageType2VolumeLayout.Items { if layout != nil { - tmp := layout.ToMap() + tmp := layout.(*VolumeLayout).ToMap() tmp["collection"] = c.Name layouts = append(layouts, tmp) } diff --git a/go/topology/topology_vacuum.go b/go/topology/topology_vacuum.go index a1d6d2564..d6fa2213e 100644 --- a/go/topology/topology_vacuum.go +++ b/go/topology/topology_vacuum.go @@ -1,13 +1,14 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "net/url" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" ) func batchVacuumVolumeCheck(vl *VolumeLayout, vid storage.VolumeId, locationlist *VolumeLocationList, garbageThreshold string) bool { @@ -79,13 +80,15 @@ func batchVacuumVolumeCommit(vl *VolumeLayout, vid storage.VolumeId, locationlis return isCommitSuccess } func (t *Topology) Vacuum(garbageThreshold string) int { - for _, c := range t.collectionMap { - for _, vl := range c.replicaType2VolumeLayout { + for _, col := range t.collectionMap.Items { + c := col.(*Collection) + for _, vl := range c.storageType2VolumeLayout.Items { if vl != nil { - for vid, locationlist := range vl.vid2location { - if batchVacuumVolumeCheck(vl, vid, locationlist, garbageThreshold) { - if batchVacuumVolumeCompact(vl, vid, locationlist) { - batchVacuumVolumeCommit(vl, vid, locationlist) + volumeLayout := vl.(*VolumeLayout) + for vid, locationlist := range volumeLayout.vid2location { + if batchVacuumVolumeCheck(volumeLayout, vid, locationlist, garbageThreshold) { + if batchVacuumVolumeCompact(volumeLayout, vid, locationlist) { + batchVacuumVolumeCommit(volumeLayout, vid, locationlist) } } } diff --git a/go/topology/volume_growth.go b/go/topology/volume_growth.go index 4965e3ba0..b1f241990 100644 --- a/go/topology/volume_growth.go +++ b/go/topology/volume_growth.go @@ -1,11 +1,12 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "fmt" "math/rand" "sync" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) /* @@ -19,6 +20,7 @@ This package is created to resolve these replica placement issues: type VolumeGrowOption struct { Collection string ReplicaPlacement *storage.ReplicaPlacement + Ttl *storage.TTL DataCenter string Rack string DataNode string @@ -184,8 +186,15 @@ func (vg *VolumeGrowth) findEmptySlotsForOneVolume(topo *Topology, option *Volum func (vg *VolumeGrowth) grow(topo *Topology, vid storage.VolumeId, option *VolumeGrowOption, servers ...*DataNode) error { for _, server := range servers { - if err := AllocateVolume(server, vid, option.Collection, option.ReplicaPlacement); err == nil { - vi := storage.VolumeInfo{Id: vid, Size: 0, Collection: option.Collection, ReplicaPlacement: option.ReplicaPlacement, Version: storage.CurrentVersion} + if err := AllocateVolume(server, vid, option); err == nil { + vi := storage.VolumeInfo{ + Id: vid, + Size: 0, + Collection: option.Collection, + ReplicaPlacement: option.ReplicaPlacement, + Ttl: option.Ttl, + Version: storage.CurrentVersion, + } server.AddOrUpdateVolume(vi) topo.RegisterVolumeLayout(vi, server) glog.V(0).Infoln("Created Volume", vid, "on", server) diff --git a/go/topology/volume_growth_test.go b/go/topology/volume_growth_test.go index d2913ef5c..267b36042 100644 --- a/go/topology/volume_growth_test.go +++ b/go/topology/volume_growth_test.go @@ -1,11 +1,12 @@ package topology import ( - "code.google.com/p/weed-fs/go/sequence" - "code.google.com/p/weed-fs/go/storage" "encoding/json" "fmt" "testing" + + "github.com/chrislusf/weed-fs/go/sequence" + "github.com/chrislusf/weed-fs/go/storage" ) var topologyLayout = ` diff --git a/go/topology/volume_layout.go b/go/topology/volume_layout.go index 538acb54c..de72bf895 100644 --- a/go/topology/volume_layout.go +++ b/go/topology/volume_layout.go @@ -1,25 +1,28 @@ package topology import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "errors" "math/rand" "sync" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) // mapping from volume to its locations, inverted from server to volume type VolumeLayout struct { rp *storage.ReplicaPlacement + ttl *storage.TTL vid2location map[storage.VolumeId]*VolumeLocationList writables []storage.VolumeId // transient array of writable volume id volumeSizeLimit uint64 accessLock sync.Mutex } -func NewVolumeLayout(rp *storage.ReplicaPlacement, volumeSizeLimit uint64) *VolumeLayout { +func NewVolumeLayout(rp *storage.ReplicaPlacement, ttl *storage.TTL, volumeSizeLimit uint64) *VolumeLayout { return &VolumeLayout{ rp: rp, + ttl: ttl, vid2location: make(map[storage.VolumeId]*VolumeLocationList), writables: *new([]storage.VolumeId), volumeSizeLimit: volumeSizeLimit, @@ -42,6 +45,14 @@ func (vl *VolumeLayout) RegisterVolume(v *storage.VolumeInfo, dn *DataNode) { } } +func (vl *VolumeLayout) UnRegisterVolume(v *storage.VolumeInfo, dn *DataNode) { + vl.accessLock.Lock() + defer vl.accessLock.Unlock() + + vl.removeFromWritable(v.Id) + delete(vl.vid2location, v.Id) +} + func (vl *VolumeLayout) AddToWritable(vid storage.VolumeId) { for _, id := range vl.writables { if vid == id { @@ -192,6 +203,7 @@ func (vl *VolumeLayout) SetVolumeCapacityFull(vid storage.VolumeId) bool { func (vl *VolumeLayout) ToMap() map[string]interface{} { m := make(map[string]interface{}) m["replication"] = vl.rp.String() + m["ttl"] = vl.ttl.String() m["writables"] = vl.writables //m["locations"] = vl.vid2location return m diff --git a/go/topology/volume_location_list.go b/go/topology/volume_location_list.go index 176f469b9..011614013 100644 --- a/go/topology/volume_location_list.go +++ b/go/topology/volume_location_list.go @@ -1,7 +1,5 @@ package topology -import () - type VolumeLocationList struct { list []*DataNode } diff --git a/go/util/bytes.go b/go/util/bytes.go index 6cc3d7018..dfa4ae665 100644 --- a/go/util/bytes.go +++ b/go/util/bytes.go @@ -1,5 +1,7 @@ package util +// big endian + func BytesToUint64(b []byte) (v uint64) { length := uint(len(b)) for i := uint(0); i < length-1; i++ { @@ -18,6 +20,12 @@ func BytesToUint32(b []byte) (v uint32) { v += uint32(b[length-1]) return } +func BytesToUint16(b []byte) (v uint16) { + v += uint16(b[0]) + v <<= 8 + v += uint16(b[1]) + return +} func Uint64toBytes(b []byte, v uint64) { for i := uint(0); i < 8; i++ { b[7-i] = byte(v >> (i * 8)) @@ -28,6 +36,10 @@ func Uint32toBytes(b []byte, v uint32) { b[3-i] = byte(v >> (i * 8)) } } +func Uint16toBytes(b []byte, v uint16) { + b[0] = byte(v >> 8) + b[1] = byte(v) +} func Uint8toBytes(b []byte, v uint8) { b[0] = byte(v) } diff --git a/go/util/concurrent_read_map.go b/go/util/concurrent_read_map.go new file mode 100644 index 000000000..d16fdbcaf --- /dev/null +++ b/go/util/concurrent_read_map.go @@ -0,0 +1,37 @@ +package util + +import "sync" + +// A mostly for read map, which can thread-safely +// initialize the map entries. +type ConcurrentReadMap struct { + rmutex sync.RWMutex + mutex sync.Mutex + Items map[string]interface{} +} + +func NewConcurrentReadMap() *ConcurrentReadMap { + return &ConcurrentReadMap{Items: make(map[string]interface{})} +} + +func (m *ConcurrentReadMap) initMapEntry(key string, newEntry func() interface{}) (value interface{}) { + m.mutex.Lock() + defer m.mutex.Unlock() + if value, ok := m.Items[key]; ok { + return value + } + value = newEntry() + m.Items[key] = value + return value +} + +func (m *ConcurrentReadMap) Get(key string, newEntry func() interface{}) interface{} { + m.rmutex.RLock() + if value, ok := m.Items[key]; ok { + m.rmutex.RUnlock() + return value + } else { + m.rmutex.RUnlock() + return m.initMapEntry(key, newEntry) + } +} diff --git a/go/util/config.go b/go/util/config.go index 9a1ac680b..4cf1d7c64 100644 --- a/go/util/config.go +++ b/go/util/config.go @@ -10,9 +10,10 @@ package util import ( "bytes" - "code.google.com/p/weed-fs/go/glog" "encoding/json" "os" + + "github.com/chrislusf/weed-fs/go/glog" ) type Config struct { diff --git a/go/util/constants.go b/go/util/constants.go index 196b4a405..d677ba44f 100644 --- a/go/util/constants.go +++ b/go/util/constants.go @@ -1,7 +1,5 @@ package util -import () - const ( - VERSION = "0.63 beta" + VERSION = "0.67" ) diff --git a/go/util/file_util.go b/go/util/file_util.go index 9f3354011..30e24f001 100644 --- a/go/util/file_util.go +++ b/go/util/file_util.go @@ -2,9 +2,10 @@ package util import ( "bufio" - "code.google.com/p/weed-fs/go/glog" "errors" "os" + + "github.com/chrislusf/weed-fs/go/glog" ) func TestFolderWritable(folder string) (err error) { diff --git a/go/util/net_timeout.go b/go/util/net_timeout.go index a6cc81c99..f274e4802 100644 --- a/go/util/net_timeout.go +++ b/go/util/net_timeout.go @@ -1,9 +1,10 @@ package util import ( - "code.google.com/p/weed-fs/go/stats" "net" "time" + + "github.com/chrislusf/weed-fs/go/stats" ) // Listener wraps a net.Listener, and gives a place to store the timeout diff --git a/go/weed/benchmark.go b/go/weed/benchmark.go index eab923751..f4f0b1874 100644 --- a/go/weed/benchmark.go +++ b/go/weed/benchmark.go @@ -2,9 +2,6 @@ package main import ( "bufio" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/util" "fmt" "io" "math" @@ -16,6 +13,10 @@ import ( "strings" "sync" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/util" ) type BenchmarkOptions struct { @@ -30,11 +31,14 @@ type BenchmarkOptions struct { sequentialRead *bool collection *string cpuprofile *string + maxCpu *int vid2server map[string]string //cache for vid locations + } var ( - b BenchmarkOptions + b BenchmarkOptions + sharedBytes []byte ) func init() { @@ -50,33 +54,35 @@ func init() { b.read = cmdBenchmark.Flag.Bool("read", true, "enable read") b.sequentialRead = cmdBenchmark.Flag.Bool("readSequentially", false, "randomly read by ids from \"-list\" specified file") b.collection = cmdBenchmark.Flag.String("collection", "benchmark", "write data to this collection") - b.cpuprofile = cmdBenchmark.Flag.String("cpuprofile", "", "write cpu profile to file") + b.cpuprofile = cmdBenchmark.Flag.String("cpuprofile", "", "cpu profile output file") + b.maxCpu = cmdBenchmark.Flag.Int("maxCpu", 0, "maximum number of CPUs. 0 means all available CPUs") b.vid2server = make(map[string]string) + sharedBytes = make([]byte, 1024) } var cmdBenchmark = &Command{ UsageLine: "benchmark -server=localhost:9333 -c=10 -n=100000", Short: "benchmark on writing millions of files and read out", Long: `benchmark on an empty weed file system. - + Two tests during benchmark: 1) write lots of small files to the system 2) read the files out - + The file content is mostly zero, but no compression is done. - + You can choose to only benchmark read or write. During write, the list of uploaded file ids is stored in "-list" specified file. You can also use your own list of file ids to run read test. - + Write speed and read speed will be collected. The numbers are used to get a sense of the system. Usually your network or the hard drive is the real bottleneck. - + Another thing to watch is whether the volumes are evenly distributed to each volume server. Because the 7 more benchmark volumes are randomly distributed to servers with free slots, it's highly possible some servers have uneven amount of - benchmark volumes. To remedy this, you can use this to grow the benchmark volumes + benchmark volumes. To remedy this, you can use this to grow the benchmark volumes before starting the benchmark command: http://localhost:9333/vol/grow?collection=benchmark&count=5 @@ -87,18 +93,17 @@ var cmdBenchmark = &Command{ } var ( - wait sync.WaitGroup - writeStats *stats - readStats *stats - serverLimitChan map[string]chan bool + wait sync.WaitGroup + writeStats *stats + readStats *stats ) -func init() { - serverLimitChan = make(map[string]chan bool) -} - func runbenchmark(cmd *Command, args []string) bool { - fmt.Printf("This is Weed File System version %s %s %s\n", util.VERSION, runtime.GOOS, runtime.GOARCH) + fmt.Printf("This is Seaweed File System version %s %s %s\n", util.VERSION, runtime.GOOS, runtime.GOARCH) + if *b.maxCpu < 1 { + *b.maxCpu = runtime.NumCPU() + } + runtime.GOMAXPROCS(*b.maxCpu) if *b.cpuprofile != "" { f, err := os.Create(*b.cpuprofile) if err != nil { @@ -122,12 +127,12 @@ func runbenchmark(cmd *Command, args []string) bool { func bench_write() { fileIdLineChan := make(chan string) finishChan := make(chan bool) - writeStats = newStats() + writeStats = newStats(*b.concurrency) idChan := make(chan int) - wait.Add(*b.concurrency) go writeFileIds(*b.idListFile, fileIdLineChan, finishChan) for i := 0; i < *b.concurrency; i++ { - go writeFiles(idChan, fileIdLineChan, writeStats) + wait.Add(1) + go writeFiles(idChan, fileIdLineChan, &writeStats.localStats[i]) } writeStats.start = time.Now() writeStats.total = *b.numberOfFiles @@ -138,28 +143,30 @@ func bench_write() { close(idChan) wait.Wait() writeStats.end = time.Now() - wait.Add(1) + wait.Add(2) finishChan <- true finishChan <- true - close(finishChan) wait.Wait() + close(finishChan) writeStats.printStats() } func bench_read() { fileIdLineChan := make(chan string) finishChan := make(chan bool) - readStats = newStats() - wait.Add(*b.concurrency) + readStats = newStats(*b.concurrency) go readFileIds(*b.idListFile, fileIdLineChan) readStats.start = time.Now() readStats.total = *b.numberOfFiles go readStats.checkProgress("Randomly Reading Benchmark", finishChan) for i := 0; i < *b.concurrency; i++ { - go readFiles(fileIdLineChan, readStats) + wait.Add(1) + go readFiles(fileIdLineChan, &readStats.localStats[i]) } wait.Wait() + wait.Add(1) finishChan <- true + wait.Wait() close(finishChan) readStats.end = time.Now() readStats.printStats() @@ -170,126 +177,102 @@ type delayedFile struct { fp *operation.FilePart } -func writeFiles(idChan chan int, fileIdLineChan chan string, s *stats) { +func writeFiles(idChan chan int, fileIdLineChan chan string, s *stat) { + defer wait.Done() delayedDeleteChan := make(chan *delayedFile, 100) var waitForDeletions sync.WaitGroup for i := 0; i < 7; i++ { + waitForDeletions.Add(1) go func() { - waitForDeletions.Add(1) + defer waitForDeletions.Done() for df := range delayedDeleteChan { - if df == nil { - break - } if df.enterTime.After(time.Now()) { time.Sleep(df.enterTime.Sub(time.Now())) } - fp := df.fp - serverLimitChan[fp.Server] <- true - if e := util.Delete("http://" + fp.Server + "/" + fp.Fid); e == nil { + if e := util.Delete("http://" + df.fp.Server + "/" + df.fp.Fid); e == nil { s.completed++ } else { s.failed++ } - <-serverLimitChan[fp.Server] } - waitForDeletions.Done() }() } - for { - if id, ok := <-idChan; ok { - start := time.Now() - fileSize := int64(*b.fileSize + rand.Intn(64)) - fp := &operation.FilePart{Reader: &FakeReader{id: uint64(id), size: fileSize}, FileSize: fileSize} - if assignResult, err := operation.Assign(*b.server, 1, "", *b.collection); err == nil { - fp.Server, fp.Fid, fp.Collection = assignResult.PublicUrl, assignResult.Fid, *b.collection - if _, ok := serverLimitChan[fp.Server]; !ok { - serverLimitChan[fp.Server] = make(chan bool, 7) - } - serverLimitChan[fp.Server] <- true - if _, err := fp.Upload(0, *b.server); err == nil { - if rand.Intn(100) < *b.deletePercentage { - s.total++ - delayedDeleteChan <- &delayedFile{time.Now().Add(time.Second), fp} - } else { - fileIdLineChan <- fp.Fid - } - s.completed++ - s.transferred += fileSize + for id := range idChan { + start := time.Now() + fileSize := int64(*b.fileSize + rand.Intn(64)) + fp := &operation.FilePart{Reader: &FakeReader{id: uint64(id), size: fileSize}, FileSize: fileSize} + if assignResult, err := operation.Assign(*b.server, 1, "", *b.collection, ""); err == nil { + fp.Server, fp.Fid, fp.Collection = assignResult.PublicUrl, assignResult.Fid, *b.collection + if _, err := fp.Upload(0, *b.server); err == nil { + if rand.Intn(100) < *b.deletePercentage { + s.total++ + delayedDeleteChan <- &delayedFile{time.Now().Add(time.Second), fp} } else { - s.failed++ - } - writeStats.addSample(time.Now().Sub(start)) - <-serverLimitChan[fp.Server] - if *cmdBenchmark.IsDebug { - fmt.Printf("writing %d file %s\n", id, fp.Fid) + fileIdLineChan <- fp.Fid } + s.completed++ + s.transferred += fileSize } else { s.failed++ - println("writing file error:", err.Error()) + fmt.Printf("Failed to write with error:%v\n", err) + } + writeStats.addSample(time.Now().Sub(start)) + if *cmdBenchmark.IsDebug { + fmt.Printf("writing %d file %s\n", id, fp.Fid) } } else { - break + s.failed++ + println("writing file error:", err.Error()) } } close(delayedDeleteChan) waitForDeletions.Wait() - wait.Done() } -func readFiles(fileIdLineChan chan string, s *stats) { - serverLimitChan := make(map[string]chan bool) +func readFiles(fileIdLineChan chan string, s *stat) { + defer wait.Done() masterLimitChan := make(chan bool, 1) - for { - if fid, ok := <-fileIdLineChan; ok { - if len(fid) == 0 { - continue - } - if fid[0] == '#' { - continue - } - if *cmdBenchmark.IsDebug { - fmt.Printf("reading file %s\n", fid) - } - parts := strings.SplitN(fid, ",", 2) - vid := parts[0] - start := time.Now() - if server, ok := b.vid2server[vid]; !ok { - masterLimitChan <- true - if _, now_ok := b.vid2server[vid]; !now_ok { - if ret, err := operation.Lookup(*b.server, vid); err == nil { - if len(ret.Locations) > 0 { - server = ret.Locations[0].PublicUrl - b.vid2server[vid] = server - } + for fid := range fileIdLineChan { + if len(fid) == 0 { + continue + } + if fid[0] == '#' { + continue + } + if *cmdBenchmark.IsDebug { + fmt.Printf("reading file %s\n", fid) + } + parts := strings.SplitN(fid, ",", 2) + vid := parts[0] + start := time.Now() + if server, ok := b.vid2server[vid]; !ok { + masterLimitChan <- true + if _, now_ok := b.vid2server[vid]; !now_ok { + if ret, err := operation.Lookup(*b.server, vid); err == nil { + if len(ret.Locations) > 0 { + server = ret.Locations[0].PublicUrl + b.vid2server[vid] = server } } - <-masterLimitChan } - if server, ok := b.vid2server[vid]; ok { - if _, ok := serverLimitChan[server]; !ok { - serverLimitChan[server] = make(chan bool, 7) - } - serverLimitChan[server] <- true - url := "http://" + server + "/" + fid - if bytesRead, err := util.Get(url); err == nil { - s.completed++ - s.transferred += int64(len(bytesRead)) - readStats.addSample(time.Now().Sub(start)) - } else { - s.failed++ - println("!!!! Failed to read from ", url, " !!!!!") - } - <-serverLimitChan[server] + <-masterLimitChan + } + if server, ok := b.vid2server[vid]; ok { + url := "http://" + server + "/" + fid + if bytesRead, err := util.Get(url); err == nil { + s.completed++ + s.transferred += int64(len(bytesRead)) + readStats.addSample(time.Now().Sub(start)) } else { s.failed++ - println("!!!! volume id ", vid, " location not found!!!!!") + fmt.Printf("Failed to read %s error:%v\n", url, err) } } else { - break + s.failed++ + println("!!!! volume id ", vid, " location not found!!!!!") } } - wait.Done() } func writeFileIds(fileName string, fileIdLineChan chan string, finishChan chan bool) { @@ -353,20 +336,28 @@ const ( // An efficient statics collecting and rendering type stats struct { - data []int - overflow []int + data []int + overflow []int + localStats []stat + start time.Time + end time.Time + total int +} +type stat struct { completed int failed int total int transferred int64 - start time.Time - end time.Time } var percentages = []int{50, 66, 75, 80, 90, 95, 98, 99, 100} -func newStats() *stats { - return &stats{data: make([]int, benchResolution), overflow: make([]int, 0)} +func newStats(n int) *stats { + return &stats{ + data: make([]int, benchResolution), + overflow: make([]int, 0), + localStats: make([]stat, n), + } } func (s *stats) addSample(d time.Duration) { @@ -387,28 +378,41 @@ func (s *stats) checkProgress(testName string, finishChan chan bool) { for { select { case <-finishChan: + wait.Done() return case t := <-ticker: - completed, transferred, taken := s.completed-lastCompleted, s.transferred-lastTransferred, t.Sub(lastTime) + completed, transferred, taken, total := 0, int64(0), t.Sub(lastTime), s.total + for _, localStat := range s.localStats { + completed += localStat.completed + transferred += localStat.transferred + total += localStat.total + } fmt.Printf("Completed %d of %d requests, %3.1f%% %3.1f/s %3.1fMB/s\n", - s.completed, s.total, float64(s.completed)*100/float64(s.total), - float64(completed)*float64(int64(time.Second))/float64(int64(taken)), - float64(transferred)*float64(int64(time.Second))/float64(int64(taken))/float64(1024*1024), + completed, total, float64(completed)*100/float64(total), + float64(completed-lastCompleted)*float64(int64(time.Second))/float64(int64(taken)), + float64(transferred-lastTransferred)*float64(int64(time.Second))/float64(int64(taken))/float64(1024*1024), ) - lastCompleted, lastTransferred, lastTime = s.completed, s.transferred, t + lastCompleted, lastTransferred, lastTime = completed, transferred, t } } } func (s *stats) printStats() { + completed, failed, transferred, total := 0, 0, int64(0), s.total + for _, localStat := range s.localStats { + completed += localStat.completed + failed += localStat.failed + transferred += localStat.transferred + total += localStat.total + } timeTaken := float64(int64(s.end.Sub(s.start))) / 1000000000 fmt.Printf("\nConcurrency Level: %d\n", *b.concurrency) fmt.Printf("Time taken for tests: %.3f seconds\n", timeTaken) - fmt.Printf("Complete requests: %d\n", s.completed) - fmt.Printf("Failed requests: %d\n", s.failed) - fmt.Printf("Total transferred: %d bytes\n", s.transferred) - fmt.Printf("Requests per second: %.2f [#/sec]\n", float64(s.completed)/timeTaken) - fmt.Printf("Transfer rate: %.2f [Kbytes/sec]\n", float64(s.transferred)/1024/timeTaken) + fmt.Printf("Complete requests: %d\n", completed) + fmt.Printf("Failed requests: %d\n", failed) + fmt.Printf("Total transferred: %d bytes\n", transferred) + fmt.Printf("Requests per second: %.2f [#/sec]\n", float64(completed)/timeTaken) + fmt.Printf("Transfer rate: %.2f [Kbytes/sec]\n", float64(transferred)/1024/timeTaken) n, sum := 0, 0 min, max := 10000000, 0 for i := 0; i < len(s.data); i++ { @@ -496,15 +500,32 @@ func (l *FakeReader) Read(p []byte) (n int, err error) { } else { n = len(p) } - for i := 0; i < n-8; i += 8 { - for s := uint(0); s < 8; s++ { - p[i] = byte(l.id >> (s * 8)) + if n >= 8 { + for i := 0; i < 8; i++ { + p[i] = byte(l.id >> uint(i*8)) } } l.size -= int64(n) return } +func (l *FakeReader) WriteTo(w io.Writer) (n int64, err error) { + size := int(l.size) + bufferSize := len(sharedBytes) + for size > 0 { + tempBuffer := sharedBytes + if size < bufferSize { + tempBuffer = sharedBytes[0:size] + } + count, e := w.Write(tempBuffer) + if e != nil { + return int64(size), e + } + size -= count + } + return l.size, nil +} + func Readln(r *bufio.Reader) ([]byte, error) { var ( isPrefix bool = true diff --git a/go/weed/compact.go b/go/weed/compact.go index 580f3f98d..71c4ea90f 100644 --- a/go/weed/compact.go +++ b/go/weed/compact.go @@ -1,8 +1,8 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) func init() { @@ -12,7 +12,7 @@ func init() { var cmdCompact = &Command{ UsageLine: "compact -dir=/tmp -volumeId=234", - Short: "run weed tool compact on volume file if corrupted", + Short: "run weed tool compact on volume file", Long: `Force an compaction to remove deleted files from volume files. The compacted .dat file is stored as .cpd file. The compacted .idx file is stored as .cpx file. @@ -33,7 +33,7 @@ func runCompact(cmd *Command, args []string) bool { } vid := storage.VolumeId(*compactVolumeId) - v, err := storage.NewVolume(*compactVolumePath, *compactVolumeCollection, vid, nil) + v, err := storage.NewVolume(*compactVolumePath, *compactVolumeCollection, vid, nil, nil) if err != nil { glog.Fatalf("Load Volume [ERROR] %s\n", err) } diff --git a/go/weed/download.go b/go/weed/download.go index 4309fe5e0..c782654f5 100644 --- a/go/weed/download.go +++ b/go/weed/download.go @@ -1,14 +1,15 @@ package main import ( - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/util" "fmt" "io" "io/ioutil" "os" "path" "strings" + + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/util" ) var ( diff --git a/go/weed/export.go b/go/weed/export.go index 3f8ff85bb..c9cc0e3fe 100644 --- a/go/weed/export.go +++ b/go/weed/export.go @@ -3,8 +3,6 @@ package main import ( "archive/tar" "bytes" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "fmt" "os" "path" @@ -12,6 +10,9 @@ import ( "strings" "text/template" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) func init() { @@ -27,7 +28,7 @@ var cmdExport = &Command{ UsageLine: "export -dir=/tmp -volumeId=234 -o=/dir/name.tar -fileNameFormat={{.Name}}", Short: "list or export files from one volume data file", Long: `List all files in a volume, or Export all files in a volume to a tar file if the output is specified. - + The format of file name in the tar file can be customized. Default is {{.Mime}}/{{.Id}}:{{.Name}}. Also available is {{.Key}}. `, @@ -36,7 +37,7 @@ var cmdExport = &Command{ var ( exportVolumePath = cmdExport.Flag.String("dir", "/tmp", "input data directory to store volume data files") exportCollection = cmdExport.Flag.String("collection", "", "the volume collection name") - exportVolumeId = cmdExport.Flag.Int("volumeId", -1, "a volume id. The volume should already exist in the dir. The volume index file should not exist.") + exportVolumeId = cmdExport.Flag.Int("volumeId", -1, "a volume id. The volume .dat and .idx files should already exist in the dir.") dest = cmdExport.Flag.String("o", "", "output tar file name, must ends with .tar, or just a \"-\" for stdout") format = cmdExport.Flag.String("fileNameFormat", defaultFnFormat, "filename format, default to {{.Mime}}/{{.Id}}:{{.Name}}") tarFh *tar.Writer @@ -100,7 +101,7 @@ func runExport(cmd *Command, args []string) bool { var version storage.Version err = storage.ScanVolumeFile(*exportVolumePath, *exportCollection, vid, func(superBlock storage.SuperBlock) error { - version = superBlock.Version + version = superBlock.Version() return nil }, true, func(n *storage.Needle, offset int64) error { nv, ok := nm.Get(n.Id) diff --git a/go/weed/filer.go b/go/weed/filer.go index d7d028d50..11154f183 100644 --- a/go/weed/filer.go +++ b/go/weed/filer.go @@ -1,13 +1,14 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" - "code.google.com/p/weed-fs/go/weed/weed_server" "net/http" "os" "strconv" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" + "github.com/chrislusf/weed-fs/go/weed/weed_server" ) var ( @@ -20,6 +21,7 @@ type FilerOptions struct { collection *string defaultReplicaPlacement *string dir *string + redirectOnRead *bool } func init() { @@ -28,7 +30,8 @@ func init() { f.collection = cmdFiler.Flag.String("collection", "", "all data will be stored in this collection") f.port = cmdFiler.Flag.Int("port", 8888, "filer server http listen port") f.dir = cmdFiler.Flag.String("dir", os.TempDir(), "directory to store meta data") - f.defaultReplicaPlacement = cmdFiler.Flag.String("defaultReplicaPlacement", "000", "Default replication type if not specified.") + f.defaultReplicaPlacement = cmdFiler.Flag.String("defaultReplicaPlacement", "000", "default replication type if not specified") + f.redirectOnRead = cmdFiler.Flag.Bool("redirectOnRead", false, "whether proxy or redirect to volume server during file GET request") } var cmdFiler = &Command{ @@ -54,16 +57,19 @@ var cmdFiler = &Command{ } func runFiler(cmd *Command, args []string) bool { + if err := util.TestFolderWritable(*f.dir); err != nil { glog.Fatalf("Check Meta Folder (-dir) Writable %s : %s", *f.dir, err) } r := http.NewServeMux() - _, nfs_err := weed_server.NewFilerServer(r, *f.port, *f.master, *f.dir, *f.collection) + _, nfs_err := weed_server.NewFilerServer(r, *f.port, *f.master, *f.dir, *f.collection, + *f.defaultReplicaPlacement, *f.redirectOnRead, + ) if nfs_err != nil { glog.Fatalf(nfs_err.Error()) } - glog.V(0).Infoln("Start Weed Filer", util.VERSION, "at port", strconv.Itoa(*f.port)) + glog.V(0).Infoln("Start Seaweed Filer", util.VERSION, "at port", strconv.Itoa(*f.port)) filerListener, e := util.NewListener( ":"+strconv.Itoa(*f.port), time.Duration(10)*time.Second, diff --git a/go/weed/fix.go b/go/weed/fix.go index 02147d796..e66075ed2 100644 --- a/go/weed/fix.go +++ b/go/weed/fix.go @@ -1,11 +1,12 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "os" "path" "strconv" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) func init() { @@ -16,7 +17,7 @@ func init() { var cmdFix = &Command{ UsageLine: "fix -dir=/tmp -volumeId=234", Short: "run weed tool fix on index file if corrupted", - Long: `Fix runs the WeedFS fix command to re-create the index .idx file. + Long: `Fix runs the SeeweedFS fix command to re-create the index .idx file. `, } diff --git a/go/weed/master.go b/go/weed/master.go index b95ca5cb1..f88964b6d 100644 --- a/go/weed/master.go +++ b/go/weed/master.go @@ -1,16 +1,17 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" - "code.google.com/p/weed-fs/go/weed/weed_server" - "github.com/gorilla/mux" "net/http" "os" "runtime" "strconv" "strings" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" + "github.com/chrislusf/weed-fs/go/weed/weed_server" + "github.com/gorilla/mux" ) func init() { @@ -29,6 +30,7 @@ var cmdMaster = &Command{ var ( mport = cmdMaster.Flag.Int("port", 9333, "http listen port") masterIp = cmdMaster.Flag.String("ip", "", "master listening ip address, default to listen on all network interfaces") + masterBindIp = cmdMaster.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to") mPublicIp = cmdMaster.Flag.String("publicIp", "", "peer accessible |") metaFolder = cmdMaster.Flag.String("mdir", os.TempDir(), "data directory to store meta data") masterPeers = cmdMaster.Flag.String("peers", "", "other master nodes in comma separated ip:port list") @@ -61,9 +63,9 @@ func runMaster(cmd *Command, args []string) bool { *volumeSizeLimitMB, *mpulse, *confFile, *defaultReplicaPlacement, *garbageThreshold, masterWhiteList, ) - listeningAddress := *masterIp + ":" + strconv.Itoa(*mport) + listeningAddress := *masterBindIp + ":" + strconv.Itoa(*mport) - glog.V(0).Infoln("Start Weed Master", util.VERSION, "at", listeningAddress) + glog.V(0).Infoln("Start Seaweed Master", util.VERSION, "at", listeningAddress) listener, e := util.NewListener(listeningAddress, time.Duration(*mTimeout)*time.Second) if e != nil { diff --git a/go/weed/mount.go b/go/weed/mount.go index 0290e0f53..66e645387 100644 --- a/go/weed/mount.go +++ b/go/weed/mount.go @@ -1,7 +1,5 @@ package main -import () - type MountOptions struct { filer *string dir *string diff --git a/go/weed/mount_std.go b/go/weed/mount_std.go index 9376b3f2e..808c6c563 100644 --- a/go/weed/mount_std.go +++ b/go/weed/mount_std.go @@ -3,19 +3,20 @@ package main import ( - "bazil.org/fuse" - "bazil.org/fuse/fs" - "code.google.com/p/weed-fs/go/filer" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/util" "fmt" "os" "runtime" + + "bazil.org/fuse" + "bazil.org/fuse/fs" + "github.com/chrislusf/weed-fs/go/filer" + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" ) func runMount(cmd *Command, args []string) bool { - fmt.Printf("This is Weed File System version %s %s %s\n", util.VERSION, runtime.GOOS, runtime.GOARCH) + fmt.Printf("This is Seaweed File System version %s %s %s\n", util.VERSION, runtime.GOOS, runtime.GOARCH) if *mountOptions.dir == "" { fmt.Printf("Please specify the mount directory via \"-dir\"") return false diff --git a/go/weed/server.go b/go/weed/server.go index 8b0372159..67bbdb370 100644 --- a/go/weed/server.go +++ b/go/weed/server.go @@ -1,10 +1,6 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" - "code.google.com/p/weed-fs/go/weed/weed_server" - "github.com/gorilla/mux" "net/http" "os" "runtime" @@ -13,6 +9,11 @@ import ( "strings" "sync" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" + "github.com/chrislusf/weed-fs/go/weed/weed_server" + "github.com/gorilla/mux" ) type ServerOptions struct { @@ -31,23 +32,24 @@ func init() { var cmdServer = &Command{ UsageLine: "server -port=8080 -dir=/tmp -volume.max=5 -ip=server_name", Short: "start a server, including volume server, and automatically elect a master server", - Long: `start both a volume server to provide storage spaces + Long: `start both a volume server to provide storage spaces and a master server to provide volume=>location mapping service and sequence number of file ids - + This is provided as a convenient way to start both volume server and master server. The servers are exactly the same as starting them separately. So other volume servers can use this embedded master server also. - + Optionally, one filer server can be started. Logically, filer servers should not be in a cluster. They run with meta data on disk, not shared. So each filer server is different. - + `, } var ( serverIp = cmdServer.Flag.String("ip", "", "ip or server name") serverPublicIp = cmdServer.Flag.String("publicIp", "", "ip or server name") + serverBindIp = cmdServer.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to") serverMaxCpu = cmdServer.Flag.Int("maxCpu", 0, "maximum number of CPUs. 0 means all available CPUs") serverTimeout = cmdServer.Flag.Int("idleTimeout", 10, "connection idle seconds") serverDataCenter = cmdServer.Flag.String("dataCenter", "", "current volume server's data center name") @@ -71,12 +73,14 @@ var ( ) func init() { - serverOptions.cpuprofile = cmdServer.Flag.String("cpuprofile", "", "write cpu profile to file") + serverOptions.cpuprofile = cmdServer.Flag.String("cpuprofile", "", "cpu profile output file") filerOptions.master = cmdServer.Flag.String("filer.master", "", "default to current master server") filerOptions.collection = cmdServer.Flag.String("filer.collection", "", "all data will be stored in this collection") filerOptions.port = cmdServer.Flag.Int("filer.port", 8888, "filer server http listen port") filerOptions.dir = cmdServer.Flag.String("filer.dir", "", "directory to store meta data, default to a 'filer' sub directory of what -mdir is specified") filerOptions.defaultReplicaPlacement = cmdServer.Flag.String("filer.defaultReplicaPlacement", "", "Default replication type if not specified during runtime.") + filerOptions.redirectOnRead = cmdServer.Flag.Bool("filer.redirectOnRead", false, "whether proxy or redirect to volume server during file GET request") + } func runServer(cmd *Command, args []string) bool { @@ -97,6 +101,10 @@ func runServer(cmd *Command, args []string) bool { } } + if *filerOptions.redirectOnRead { + *isStartingFiler = true + } + *filerOptions.master = *serverPublicIp + ":" + strconv.Itoa(*masterPort) if *filerOptions.defaultReplicaPlacement == "" { @@ -148,11 +156,13 @@ func runServer(cmd *Command, args []string) bool { if *isStartingFiler { go func() { r := http.NewServeMux() - _, nfs_err := weed_server.NewFilerServer(r, *filerOptions.port, *filerOptions.master, *filerOptions.dir, *filerOptions.collection) + _, nfs_err := weed_server.NewFilerServer(r, *filerOptions.port, *filerOptions.master, *filerOptions.dir, *filerOptions.collection, + *filerOptions.defaultReplicaPlacement, *filerOptions.redirectOnRead, + ) if nfs_err != nil { glog.Fatalf(nfs_err.Error()) } - glog.V(0).Infoln("Start Weed Filer", util.VERSION, "at port", strconv.Itoa(*filerOptions.port)) + glog.V(0).Infoln("Start Seaweed Filer", util.VERSION, "at port", strconv.Itoa(*filerOptions.port)) filerListener, e := util.NewListener( ":"+strconv.Itoa(*filerOptions.port), time.Duration(10)*time.Second, @@ -178,8 +188,8 @@ func runServer(cmd *Command, args []string) bool { *masterVolumeSizeLimitMB, *volumePulse, *masterConfFile, *masterDefaultReplicaPlacement, *serverGarbageThreshold, serverWhiteList, ) - glog.V(0).Infoln("Start Weed Master", util.VERSION, "at", *serverIp+":"+strconv.Itoa(*masterPort)) - masterListener, e := util.NewListener(*serverIp+":"+strconv.Itoa(*masterPort), time.Duration(*serverTimeout)*time.Second) + glog.V(0).Infoln("Start Seaweed Master", util.VERSION, "at", *serverIp+":"+strconv.Itoa(*masterPort)) + masterListener, e := util.NewListener(*serverBindIp+":"+strconv.Itoa(*masterPort), time.Duration(*serverTimeout)*time.Second) if e != nil { glog.Fatalf(e.Error()) } @@ -211,9 +221,9 @@ func runServer(cmd *Command, args []string) bool { *volumeFixJpgOrientation, ) - glog.V(0).Infoln("Start Weed volume server", util.VERSION, "at", *serverIp+":"+strconv.Itoa(*volumePort)) + glog.V(0).Infoln("Start Seaweed volume server", util.VERSION, "at", *serverIp+":"+strconv.Itoa(*volumePort)) volumeListener, e := util.NewListener( - *serverIp+":"+strconv.Itoa(*volumePort), + *serverBindIp+":"+strconv.Itoa(*volumePort), time.Duration(*serverTimeout)*time.Second, ) if e != nil { diff --git a/go/weed/shell.go b/go/weed/shell.go index 6a3331284..f2c4990ea 100644 --- a/go/weed/shell.go +++ b/go/weed/shell.go @@ -2,9 +2,10 @@ package main import ( "bufio" - "code.google.com/p/weed-fs/go/glog" "fmt" "os" + + "github.com/chrislusf/weed-fs/go/glog" ) func init() { diff --git a/go/weed/signal_handling_notsupported.go b/go/weed/signal_handling_notsupported.go index ad4f37b0c..343cf7de2 100644 --- a/go/weed/signal_handling_notsupported.go +++ b/go/weed/signal_handling_notsupported.go @@ -2,7 +2,5 @@ package main -import () - func OnInterrupt(fn func()) { } diff --git a/go/weed/upload.go b/go/weed/upload.go index b59313a2a..2d67c0bd9 100644 --- a/go/weed/upload.go +++ b/go/weed/upload.go @@ -1,17 +1,19 @@ package main import ( - "code.google.com/p/weed-fs/go/operation" "encoding/json" "fmt" "os" "path/filepath" + + "github.com/chrislusf/weed-fs/go/operation" ) var ( uploadReplication *string uploadCollection *string uploadDir *string + uploadTtl *string include *string maxMB *int ) @@ -24,6 +26,7 @@ func init() { include = cmdUpload.Flag.String("include", "", "pattens of files to upload, e.g., *.pdf, *.html, ab?d.txt, works together with -dir") uploadReplication = cmdUpload.Flag.String("replication", "", "replication type") uploadCollection = cmdUpload.Flag.String("collection", "", "optional collection name") + uploadTtl = cmdUpload.Flag.String("ttl", "", "time to live, e.g.: 1m, 1h, 1d, 1M, 1y") maxMB = cmdUpload.Flag.Int("maxMB", 0, "split files larger than the limit") } @@ -67,7 +70,7 @@ func runUpload(cmd *Command, args []string) bool { if e != nil { return e } - results, e := operation.SubmitFiles(*server, parts, *uploadReplication, *uploadCollection, *maxMB) + results, e := operation.SubmitFiles(*server, parts, *uploadReplication, *uploadCollection, *uploadTtl, *maxMB) bytes, _ := json.Marshal(results) fmt.Println(string(bytes)) if e != nil { @@ -84,7 +87,7 @@ func runUpload(cmd *Command, args []string) bool { if e != nil { fmt.Println(e.Error()) } - results, _ := operation.SubmitFiles(*server, parts, *uploadReplication, *uploadCollection, *maxMB) + results, _ := operation.SubmitFiles(*server, parts, *uploadReplication, *uploadCollection, *uploadTtl, *maxMB) bytes, _ := json.Marshal(results) fmt.Println(string(bytes)) } diff --git a/go/weed/version.go b/go/weed/version.go index 2e0a59822..8d3a6fed7 100644 --- a/go/weed/version.go +++ b/go/weed/version.go @@ -1,16 +1,17 @@ package main import ( - "code.google.com/p/weed-fs/go/util" "fmt" "runtime" + + "github.com/chrislusf/weed-fs/go/util" ) var cmdVersion = &Command{ Run: runVersion, UsageLine: "version", - Short: "print Weed File System version", - Long: `Version prints the Weed File System version`, + Short: "print Seaweed File System version", + Long: `Version prints the Seaweed File System version`, } func runVersion(cmd *Command, args []string) bool { diff --git a/go/weed/volume.go b/go/weed/volume.go index 8aa52c43d..212cb4b33 100644 --- a/go/weed/volume.go +++ b/go/weed/volume.go @@ -1,15 +1,16 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/util" - "code.google.com/p/weed-fs/go/weed/weed_server" "net/http" "os" "runtime" "strconv" "strings" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/util" + "github.com/chrislusf/weed-fs/go/weed/weed_server" ) func init() { @@ -30,6 +31,7 @@ var ( maxVolumeCounts = cmdVolume.Flag.String("max", "7", "maximum numbers of volumes, count[,count]...") ip = cmdVolume.Flag.String("ip", "", "ip or server name") publicIp = cmdVolume.Flag.String("publicIp", "", "Publicly accessible ") + volumeBindIp = cmdVolume.Flag.String("ip.bind", "0.0.0.0", "ip address to bind to") masterNode = cmdVolume.Flag.String("mserver", "localhost:9333", "master server location") vpulse = cmdVolume.Flag.Int("pulseSeconds", 5, "number of seconds between heartbeats, must be smaller than or equal to the master's setting") vTimeout = cmdVolume.Flag.Int("idleTimeout", 10, "connection idle seconds") @@ -84,9 +86,9 @@ func runVolume(cmd *Command, args []string) bool { *fixJpgOrientation, ) - listeningAddress := *ip + ":" + strconv.Itoa(*vport) + listeningAddress := *volumeBindIp + ":" + strconv.Itoa(*vport) - glog.V(0).Infoln("Start Weed volume server", util.VERSION, "at", listeningAddress) + glog.V(0).Infoln("Start Seaweed volume server", util.VERSION, "at", listeningAddress) listener, e := util.NewListener(listeningAddress, time.Duration(*vTimeout)*time.Second) if e != nil { diff --git a/go/weed/volume_test.go b/go/weed/volume_test.go index 2499d8543..ef00a8c7c 100644 --- a/go/weed/volume_test.go +++ b/go/weed/volume_test.go @@ -1,10 +1,11 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" "net/http" "testing" "time" + + "github.com/chrislusf/weed-fs/go/glog" ) func TestXYZ(t *testing.T) { diff --git a/go/weed/weed.go b/go/weed/weed.go index c739d8e93..c304b7f35 100644 --- a/go/weed/weed.go +++ b/go/weed/weed.go @@ -1,7 +1,6 @@ package main import ( - "code.google.com/p/weed-fs/go/glog" "flag" "fmt" "io" @@ -13,6 +12,8 @@ import ( "time" "unicode" "unicode/utf8" + + "github.com/chrislusf/weed-fs/go/glog" ) var IsDebug *bool @@ -90,7 +91,7 @@ func main() { } var usageTemplate = ` -Weed File System : store billions of files and serve them fast! +Seaweed File System : store billions of files and serve them fast! Usage: diff --git a/go/weed/weed_server/common.go b/go/weed/weed_server/common.go index a547d7462..39b0830e3 100644 --- a/go/weed/weed_server/common.go +++ b/go/weed/weed_server/common.go @@ -2,11 +2,6 @@ package weed_server import ( "bytes" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/stats" - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/util" "encoding/json" "fmt" "net" @@ -14,6 +9,12 @@ import ( "path/filepath" "strconv" "strings" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/stats" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/util" ) var serverStats *stats.ServerStats @@ -99,14 +100,14 @@ func submitForClientHandler(w http.ResponseWriter, r *http.Request, masterUrl st } debug("parsing upload file...") - fname, data, mimeType, isGzipped, lastModified, pe := storage.ParseUpload(r) + fname, data, mimeType, isGzipped, lastModified, _, pe := storage.ParseUpload(r) if pe != nil { writeJsonError(w, r, pe) return } debug("assigning file id for", fname) - assignResult, ae := operation.Assign(masterUrl, 1, r.FormValue("replication"), r.FormValue("collection")) + assignResult, ae := operation.Assign(masterUrl, 1, r.FormValue("replication"), r.FormValue("collection"), r.FormValue("ttl")) if ae != nil { writeJsonError(w, r, ae) return diff --git a/go/weed/weed_server/filer_server.go b/go/weed/weed_server/filer_server.go index e56cb5964..0bda58d06 100644 --- a/go/weed/weed_server/filer_server.go +++ b/go/weed/weed_server/filer_server.go @@ -1,24 +1,31 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/filer" - "code.google.com/p/weed-fs/go/glog" "net/http" "strconv" + + "github.com/chrislusf/weed-fs/go/filer" + "github.com/chrislusf/weed-fs/go/glog" ) type FilerServer struct { - port string - master string - collection string - filer filer.Filer + port string + master string + collection string + defaultReplication string + redirectOnRead bool + filer filer.Filer } -func NewFilerServer(r *http.ServeMux, port int, master string, dir string, collection string) (fs *FilerServer, err error) { +func NewFilerServer(r *http.ServeMux, port int, master string, dir string, collection string, + replication string, redirectOnRead bool, +) (fs *FilerServer, err error) { fs = &FilerServer{ - master: master, - collection: collection, - port: ":" + strconv.Itoa(port), + master: master, + collection: collection, + defaultReplication: replication, + redirectOnRead: redirectOnRead, + port: ":" + strconv.Itoa(port), } if fs.filer, err = filer.NewFilerEmbedded(master, dir); err != nil { diff --git a/go/weed/weed_server/filer_server_handlers.go b/go/weed/weed_server/filer_server_handlers.go index ab30aaaed..6f22912a7 100644 --- a/go/weed/weed_server/filer_server_handlers.go +++ b/go/weed/weed_server/filer_server_handlers.go @@ -1,12 +1,8 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" - "github.com/syndtr/goleveldb/leveldb" "io" "io/ioutil" "math/rand" @@ -14,6 +10,11 @@ import ( "net/url" "strconv" "strings" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/util" + "github.com/syndtr/goleveldb/leveldb" ) func (fs *FilerServer) filerHandler(w http.ResponseWriter, r *http.Request) { @@ -80,7 +81,12 @@ func (fs *FilerServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request, return } urlLocation := lookup.Locations[rand.Intn(len(lookup.Locations))].PublicUrl - u, _ := url.Parse("http://" + urlLocation + "/" + fileId) + urlString := "http://" + urlLocation + "/" + fileId + if fs.redirectOnRead { + http.Redirect(w, r, urlString, http.StatusFound) + return + } + u, _ := url.Parse(urlString) request := &http.Request{ Method: r.Method, URL: u, @@ -103,12 +109,17 @@ func (fs *FilerServer) GetOrHeadHandler(w http.ResponseWriter, r *http.Request, for k, v := range resp.Header { w.Header()[k] = v } + w.WriteHeader(resp.StatusCode) io.Copy(w, resp.Body) } func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) { query := r.URL.Query() - assignResult, ae := operation.Assign(fs.master, 1, query.Get("replication"), fs.collection) + replication := query.Get("replication") + if replication == "" { + replication = fs.defaultReplication + } + assignResult, ae := operation.Assign(fs.master, 1, replication, fs.collection, query.Get("ttl")) if ae != nil { glog.V(0).Infoln("failing to assign a file id", ae.Error()) writeJsonError(w, r, ae) @@ -130,14 +141,14 @@ func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) { } resp, do_err := util.Do(request) if do_err != nil { - glog.V(0).Infoln("failing to connect to volume server", do_err.Error()) + glog.V(0).Infoln("failing to connect to volume server", r.RequestURI, do_err.Error()) writeJsonError(w, r, do_err) return } defer resp.Body.Close() resp_body, ra_err := ioutil.ReadAll(resp.Body) if ra_err != nil { - glog.V(0).Infoln("failing to upload to volume server", ra_err.Error()) + glog.V(0).Infoln("failing to upload to volume server", r.RequestURI, ra_err.Error()) writeJsonError(w, r, ra_err) return } @@ -145,12 +156,12 @@ func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) { var ret operation.UploadResult unmarshal_err := json.Unmarshal(resp_body, &ret) if unmarshal_err != nil { - glog.V(0).Infoln("failing to read upload resonse", string(resp_body)) + glog.V(0).Infoln("failing to read upload resonse", r.RequestURI, string(resp_body)) writeJsonError(w, r, unmarshal_err) return } if ret.Error != "" { - glog.V(0).Infoln("failing to post to volume server", ret.Error) + glog.V(0).Infoln("failing to post to volume server", r.RequestURI, ret.Error) writeJsonError(w, r, errors.New(ret.Error)) return } @@ -168,7 +179,7 @@ func (fs *FilerServer) PostHandler(w http.ResponseWriter, r *http.Request) { glog.V(4).Infoln("saving", path, "=>", assignResult.Fid) if db_err := fs.filer.CreateFile(path, assignResult.Fid); db_err != nil { operation.DeleteFile(fs.master, assignResult.Fid) //clean up - glog.V(0).Infoln("failing to write to filer server", db_err.Error()) + glog.V(0).Infoln("failing to write to filer server", r.RequestURI, db_err.Error()) writeJsonError(w, r, db_err) return } diff --git a/go/weed/weed_server/filer_server_handlers_admin.go b/go/weed/weed_server/filer_server_handlers_admin.go index b4b8b7221..5ba12e0b8 100644 --- a/go/weed/weed_server/filer_server_handlers_admin.go +++ b/go/weed/weed_server/filer_server_handlers_admin.go @@ -1,8 +1,9 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" "net/http" + + "github.com/chrislusf/weed-fs/go/glog" ) /* diff --git a/go/weed/weed_server/master_server.go b/go/weed/weed_server/master_server.go index a89959fa4..d000cb610 100644 --- a/go/weed/weed_server/master_server.go +++ b/go/weed/weed_server/master_server.go @@ -1,16 +1,17 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/sequence" - "code.google.com/p/weed-fs/go/topology" - "code.google.com/p/weed-fs/go/util" - "github.com/goraft/raft" - "github.com/gorilla/mux" "net/http" "net/http/httputil" "net/url" "sync" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/sequence" + "github.com/chrislusf/weed-fs/go/topology" + "github.com/chrislusf/weed-fs/go/util" + "github.com/goraft/raft" + "github.com/gorilla/mux" ) type MasterServer struct { diff --git a/go/weed/weed_server/master_server_handlers.go b/go/weed/weed_server/master_server_handlers.go index 1b58e6e73..7a7a3b70d 100644 --- a/go/weed/weed_server/master_server_handlers.go +++ b/go/weed/weed_server/master_server_handlers.go @@ -1,12 +1,13 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/stats" - "code.google.com/p/weed-fs/go/storage" "net/http" "strconv" "strings" + + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/stats" + "github.com/chrislusf/weed-fs/go/storage" ) func (ms *MasterServer) lookupVolumeId(vids []string, collection string) (volumeLocations map[string]operation.LookupResult) { @@ -78,7 +79,7 @@ func (ms *MasterServer) dirAssignHandler(w http.ResponseWriter, r *http.Request) return } - if !ms.Topo.HasWriableVolume(option) { + if !ms.Topo.HasWritableVolume(option) { if ms.Topo.FreeSpace() <= 0 { w.WriteHeader(http.StatusNotFound) writeJsonQuiet(w, r, operation.AssignResult{Error: "No free volumes left!"}) @@ -86,7 +87,7 @@ func (ms *MasterServer) dirAssignHandler(w http.ResponseWriter, r *http.Request) } else { ms.vgLock.Lock() defer ms.vgLock.Unlock() - if !ms.Topo.HasWriableVolume(option) { + if !ms.Topo.HasWritableVolume(option) { if _, err = ms.vg.AutomaticGrowByType(option, ms.Topo); err != nil { writeJsonQuiet(w, r, operation.AssignResult{Error: "Cannot grow volume group! " + err.Error()}) return diff --git a/go/weed/weed_server/master_server_handlers_admin.go b/go/weed/weed_server/master_server_handlers_admin.go index d50075fd5..d7124e567 100644 --- a/go/weed/weed_server/master_server_handlers_admin.go +++ b/go/weed/weed_server/master_server_handlers_admin.go @@ -1,18 +1,19 @@ package weed_server import ( - proto "code.google.com/p/goprotobuf/proto" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/topology" - "code.google.com/p/weed-fs/go/util" "encoding/json" "errors" "io/ioutil" "net/http" "strconv" "strings" + + proto "code.google.com/p/goprotobuf/proto" + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/topology" + "github.com/chrislusf/weed-fs/go/util" ) func (ms *MasterServer) collectionDeleteHandler(w http.ResponseWriter, r *http.Request) { @@ -55,7 +56,7 @@ func (ms *MasterServer) dirJoinHandler(w http.ResponseWriter, r *http.Request) { } } - ms.Topo.RegisterVolumes(joinMessage) + ms.Topo.ProcessJoinMessage(joinMessage) writeJsonQuiet(w, r, operation.JoinResult{VolumeSizeLimit: uint64(ms.volumeSizeLimitMB) * 1024 * 1024}) } @@ -143,8 +144,8 @@ func (ms *MasterServer) deleteFromMasterServerHandler(w http.ResponseWriter, r * } } -func (ms *MasterServer) hasWriableVolume(option *topology.VolumeGrowOption) bool { - vl := ms.Topo.GetVolumeLayout(option.Collection, option.ReplicaPlacement) +func (ms *MasterServer) HasWritableVolume(option *topology.VolumeGrowOption) bool { + vl := ms.Topo.GetVolumeLayout(option.Collection, option.ReplicaPlacement, option.Ttl) return vl.GetActiveVolumeCount(option) > 0 } @@ -157,9 +158,14 @@ func (ms *MasterServer) getVolumeGrowOption(r *http.Request) (*topology.VolumeGr if err != nil { return nil, err } + ttl, err := storage.ReadTTL(r.FormValue("ttl")) + if err != nil { + return nil, err + } volumeGrowOption := &topology.VolumeGrowOption{ Collection: r.FormValue("collection"), ReplicaPlacement: replicaPlacement, + Ttl: ttl, DataCenter: r.FormValue("dataCenter"), Rack: r.FormValue("rack"), DataNode: r.FormValue("dataNode"), diff --git a/go/weed/weed_server/raft_server.go b/go/weed/weed_server/raft_server.go index f67caaebd..b9aaef2b0 100644 --- a/go/weed/weed_server/raft_server.go +++ b/go/weed/weed_server/raft_server.go @@ -2,19 +2,22 @@ package weed_server import ( "bytes" - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/topology" "encoding/json" "errors" "fmt" - "github.com/goraft/raft" - "github.com/gorilla/mux" "io/ioutil" "math/rand" "net/http" "net/url" + "os" + "path" "strings" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/topology" + "github.com/goraft/raft" + "github.com/gorilla/mux" ) type RaftServer struct { @@ -44,6 +47,14 @@ func NewRaftServer(r *mux.Router, peers []string, httpAddr string, dataDir strin var err error transporter := raft.NewHTTPTransporter("/cluster", 0) transporter.Transport.MaxIdleConnsPerHost = 1024 + glog.V(1).Infof("Starting RaftServer with IP:%v:", httpAddr) + + // Clear old cluster configurations if peers are set + if len(s.peers) > 0 { + os.RemoveAll(path.Join(s.dataDir, "conf")) + os.RemoveAll(path.Join(s.dataDir, "log")) + os.RemoveAll(path.Join(s.dataDir, "snapshot")) + } s.raftServer, err = raft.NewServer(s.httpAddr, s.dataDir, transporter, nil, topo, "") if err != nil { @@ -52,35 +63,30 @@ func NewRaftServer(r *mux.Router, peers []string, httpAddr string, dataDir strin } transporter.Install(s.raftServer, s) s.raftServer.SetHeartbeatInterval(1 * time.Second) - s.raftServer.SetElectionTimeout(time.Duration(pulseSeconds) * 1150 * time.Millisecond) + s.raftServer.SetElectionTimeout(time.Duration(pulseSeconds) * 3450 * time.Millisecond) s.raftServer.Start() s.router.HandleFunc("/cluster/join", s.joinHandler).Methods("POST") s.router.HandleFunc("/cluster/status", s.statusHandler).Methods("GET") - // Join to leader if specified. if len(s.peers) > 0 { - if !s.raftServer.IsLogEmpty() { - glog.V(0).Infoln("Starting cluster with existing logs.") - } else { - glog.V(0).Infoln("Joining cluster:", strings.Join(s.peers, ",")) - time.Sleep(time.Duration(rand.Intn(1000)) * time.Millisecond) - firstJoinError := s.Join(s.peers) - if firstJoinError != nil { - glog.V(0).Infoln("No existing server found. Starting as leader in the new cluster.") - _, err := s.raftServer.Do(&raft.DefaultJoinCommand{ - Name: s.raftServer.Name(), - ConnectionString: "http://" + s.httpAddr, - }) - if err != nil { - glog.V(0).Infoln(err) - return nil - } + // Join to leader if specified. + glog.V(0).Infoln("Joining cluster:", strings.Join(s.peers, ",")) + time.Sleep(time.Duration(rand.Intn(1000)) * time.Millisecond) + firstJoinError := s.Join(s.peers) + if firstJoinError != nil { + glog.V(0).Infoln("No existing server found. Starting as leader in the new cluster.") + _, err := s.raftServer.Do(&raft.DefaultJoinCommand{ + Name: s.raftServer.Name(), + ConnectionString: "http://" + s.httpAddr, + }) + if err != nil { + glog.V(0).Infoln(err) + return nil } } - - // Initialize the server by joining itself. } else if s.raftServer.IsLogEmpty() { + // Initialize the server by joining itself. glog.V(0).Infoln("Initializing new cluster") _, err := s.raftServer.Do(&raft.DefaultJoinCommand{ @@ -94,7 +100,7 @@ func NewRaftServer(r *mux.Router, peers []string, httpAddr string, dataDir strin } } else { - glog.V(0).Infoln("Recovered from log") + glog.V(0).Infoln("Old conf,log,snapshot should have been removed.") } return s diff --git a/go/weed/weed_server/raft_server_handlers.go b/go/weed/weed_server/raft_server_handlers.go index 1ce24a963..b466d9afa 100644 --- a/go/weed/weed_server/raft_server_handlers.go +++ b/go/weed/weed_server/raft_server_handlers.go @@ -1,13 +1,14 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/operation" "encoding/json" - "github.com/goraft/raft" "io/ioutil" "net/http" "strings" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/goraft/raft" ) // Handles incoming RAFT joins. diff --git a/go/weed/weed_server/volume_server.go b/go/weed/weed_server/volume_server.go index b118e8c35..0a65fd2f6 100644 --- a/go/weed/weed_server/volume_server.go +++ b/go/weed/weed_server/volume_server.go @@ -1,12 +1,13 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/storage" "math/rand" "net/http" "strconv" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/storage" ) type VolumeServer struct { @@ -35,7 +36,6 @@ func NewVolumeServer(r *http.ServeMux, ip string, port int, publicIp string, fol } vs.store = storage.NewStore(port, ip, publicUrl, folders, maxCounts) - r.HandleFunc("/submit", secure(vs.whiteList, vs.submitFromVolumeServerHandler)) r.HandleFunc("/status", secure(vs.whiteList, vs.statusHandler)) r.HandleFunc("/admin/assign_volume", secure(vs.whiteList, vs.assignVolumeHandler)) r.HandleFunc("/admin/vacuum_volume_check", secure(vs.whiteList, vs.vacuumVolumeCheckHandler)) diff --git a/go/weed/weed_server/volume_server_handlers.go b/go/weed/weed_server/volume_server_handlers.go index eed198e4a..83f614941 100644 --- a/go/weed/weed_server/volume_server_handlers.go +++ b/go/weed/weed_server/volume_server_handlers.go @@ -1,12 +1,6 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/images" - "code.google.com/p/weed-fs/go/operation" - "code.google.com/p/weed-fs/go/stats" - "code.google.com/p/weed-fs/go/storage" - "code.google.com/p/weed-fs/go/topology" "io" "mime" "mime/multipart" @@ -14,6 +8,13 @@ import ( "strconv" "strings" "time" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/images" + "github.com/chrislusf/weed-fs/go/operation" + "github.com/chrislusf/weed-fs/go/stats" + "github.com/chrislusf/weed-fs/go/storage" + "github.com/chrislusf/weed-fs/go/topology" ) var fileNameEscaper = strings.NewReplacer("\\", "\\\\", "\"", "\\\"") diff --git a/go/weed/weed_server/volume_server_handlers_admin.go b/go/weed/weed_server/volume_server_handlers_admin.go index 5921a3b96..caf4c3be8 100644 --- a/go/weed/weed_server/volume_server_handlers_admin.go +++ b/go/weed/weed_server/volume_server_handlers_admin.go @@ -1,11 +1,12 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" - "code.google.com/p/weed-fs/go/stats" - "code.google.com/p/weed-fs/go/util" "net/http" "path/filepath" + + "github.com/chrislusf/weed-fs/go/glog" + "github.com/chrislusf/weed-fs/go/stats" + "github.com/chrislusf/weed-fs/go/util" ) func (vs *VolumeServer) statusHandler(w http.ResponseWriter, r *http.Request) { @@ -16,7 +17,7 @@ func (vs *VolumeServer) statusHandler(w http.ResponseWriter, r *http.Request) { } func (vs *VolumeServer) assignVolumeHandler(w http.ResponseWriter, r *http.Request) { - err := vs.store.AddVolume(r.FormValue("volume"), r.FormValue("collection"), r.FormValue("replication")) + err := vs.store.AddVolume(r.FormValue("volume"), r.FormValue("collection"), r.FormValue("replication"), r.FormValue("ttl")) if err == nil { writeJsonQuiet(w, r, map[string]string{"error": ""}) } else { @@ -50,10 +51,6 @@ func (vs *VolumeServer) freezeVolumeHandler(w http.ResponseWriter, r *http.Reque glog.V(2).Infoln("freeze volume =", r.FormValue("volume"), ", error =", err) } -func (vs *VolumeServer) submitFromVolumeServerHandler(w http.ResponseWriter, r *http.Request) { - submitForClientHandler(w, r, vs.masterNode) -} - func (vs *VolumeServer) statsDiskHandler(w http.ResponseWriter, r *http.Request) { m := make(map[string]interface{}) m["Version"] = util.VERSION diff --git a/go/weed/weed_server/volume_server_handlers_vacuum.go b/go/weed/weed_server/volume_server_handlers_vacuum.go index 60a5e9742..f115e3b8b 100644 --- a/go/weed/weed_server/volume_server_handlers_vacuum.go +++ b/go/weed/weed_server/volume_server_handlers_vacuum.go @@ -1,8 +1,9 @@ package weed_server import ( - "code.google.com/p/weed-fs/go/glog" "net/http" + + "github.com/chrislusf/weed-fs/go/glog" ) func (vs *VolumeServer) vacuumVolumeCheckHandler(w http.ResponseWriter, r *http.Request) {