chrislu
50c5dd7313
test low active brokers
10 months ago
chrislu
7598922b41
assign followers
10 months ago
chrislu
7e443ef0a1
latest protoc-gen-go
10 months ago
chrislu
ead27a4f59
publish init add followers
10 months ago
chrislu
1b4484bf0a
go fmt
10 months ago
Sean Ross
0833057503
Moved noapplexattr under runtime.GOARCH == "amd64" to resolve the fin… ( #5351 )
Moved noapplexattr under runtime.GOARCH == "amd64" to resolve the finder copy bug on arm64 macOS devices.
10 months ago
jerebear12
85d6d5371b
Disable filer UI in configuration ( #5297 )
* Add filer.ui.enabled configuration property
* Add filer.expose_directory_metadata to config
* Ammend commit
* Remove ShowUI reference
* Update all routes that allow directory metadata
* Add cmd flag to server.go
10 months ago
chrislu
54d7748a4a
3.63
10 months ago
Konstantin Lebedev
7187346cc1
avoid unexpected compact size ( #5272 )
https://github.com/seaweedfs/seaweedfs/issues/5215
10 months ago
chrislu
7c45992c79
when verifyNeedleIntegrity, do not truncate files
related to https://github.com/seaweedfs/seaweedfs/pull/5259
10 months ago
Chris Lu
32cb9b3276
fix unaligned 64-bit atomic operation in weed mount on arm when readi… ( #5308 )
fix unaligned 64-bit atomic operation in weed mount on arm when reading a file
fix #5306
10 months ago
7y-9
41ec9d7759
refactor: put the auth outside ( #5313 )
10 months ago
7y-9
f037c09c11
fix: only admin auth can delete S3 bucket ( #5312 )
10 months ago
Konstantin Lebedev
6181aa7594
fix: fs verify error counter ( #5261 )
10 months ago
Sébastien
c694cb4e7d
filer: reduce delete entry ErrNotFound logs ( #5301 )
10 months ago
Konstantin Lebedev
8fb978d6f6
[volume] requests that were canceled by the client are not an internal server problem(500 => 499) ( #5304 )
10 months ago
Justin Dhillon
a48e2ec45b
Fix Broken Links ( #5287 )
* https://learn.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-setprocessworkingsetsize
* https://learn.microsoft.com/en-us/windows/win32/api/memoryapi/nf-memoryapi-getprocessworkingsetsize
* remove https://github.com/h5bp/html5-boilerplate/blob/master/src/css/main.css
* https://github.com/AShiou/hof
10 months ago
Chris Lu
16b4f9d9ab
subscriber find broker leader first
11 months ago
Chris Lu
e27deed4bc
upgrade protoc
11 months ago
Chris Lu
656b78d1af
proxy to broker leader
11 months ago
Chris Lu
90fcde0e26
report EOF error
11 months ago
Chris Lu
07028a4864
minor
11 months ago
Chris Lu
fa72761bb8
fix on finding lock owner
11 months ago
Chris Lu
4dc6681833
proxy requests to lock owner
11 months ago
sxlehua
ffe5f19aa0
Add s3 postpolicy support header ( #5280 )
* Add s3 postpolicy support header
* optimized code
11 months ago
Chris Lu
0d74ac9224
fix divide by zero
11 months ago
Chris Lu
f9131de43f
fix lock TTL
11 months ago
Chris Lu
44b49bec99
less verbose
11 months ago
Chris Lu
56df44845f
fix when two consecutive chunks with the same modified_ts_ns
fix https://github.com/seaweedfs/seaweedfs/issues/5276
11 months ago
Chris Lu
0a12301b3d
avoid too large expiration time
11 months ago
sxlehua
0b49c16361
Adapt S3 POST ContentType ( #5275 )
11 months ago
Chris Lu
1b5ba4190c
avoid timestamp overflow
11 months ago
chrislu
d41792461c
lock returns host and owner
11 months ago
chrislu
d30150dde1
use a short-lived lock
11 months ago
chrislu
c1bb76d709
fix nil error
11 months ago
chrislu
ac50d8a822
re-lock if the lock owner is not found
11 months ago
Benoît Knecht
56287bd07d
weed/shell: Cluster check other disk types ( #5245 )
* week/shell: Cluster check other disk types
The `cluster.check` command only took the empty (`""`) and `hdd` disk types
into consideration, but a cluster with only `ssd` or `nvme` disk types would be
equally valid.
This commit simply checks that _any_ disk type is defined, and that some
volumes are available for it.
Signed-off-by: Benoît Knecht <bknecht@protonmail.ch>
* weed/shell: Replace loop that copies slice
Use the following construct instead of a `for` loop:
```golang
x = append(x, y...)
```
See https://staticcheck.dev/docs/checks#S1011 .
Signed-off-by: Benoît Knecht <bknecht@protonmail.ch>
* weed/shell: Check disk types when filer is in use
Filer stores its metadata logs in generic (i.e. `""`) or HDD disk type volumes,
so make sure those disk types exist and have volumes associated with them when
Filer is deployed in the cluster.
Signed-off-by: Benoît Knecht <bknecht@protonmail.ch>
---------
Signed-off-by: Benoît Knecht <bknecht@protonmail.ch>
11 months ago
Sébastien
0775d05a23
fix: http range request return status 500 ( #5251 )
When volume server unavailable for at least one chunk; was returning status 206.
Split `StreamContent` in two parts,
- first prepare, to get chunk info and return stream function
- then write chunk, with that stream function
That allow to catch error in first step before setting response status code in `processRangeRequest`
11 months ago
chrislu
0b2e5ddc7c
wait 3 seconds before shutting down publish client, to wait for all messages to be received
11 months ago
chrislu
545d5fbdf6
unused code
11 months ago
chrislu
a507069d19
lock for creating local partition
11 months ago
chrislu
9e6ea80dff
read existing topic conf
11 months ago
chrislu
a297247137
wait for each publish clients
11 months ago
chrislu
b6c5e57c30
read will block if no items
11 months ago
chrislu
0bf5424a2e
adjust visibility
11 months ago
chrislu
dedfd31dfb
refactor
11 months ago
chrislu
cbf750a31f
refactor
11 months ago
chrislu
b51dfe2bff
wait for publishing clients
11 months ago
chrislu
f8c55f101e
clean up
11 months ago
chrislu
d40225ecd2
return the lookup results
11 months ago