Chris Lu
e2c7be3a2e
adjust the needle mapper
as discussed on
https://github.com/chrislusf/seaweedfs/pull/642#issuecomment-387606960
7 years ago
WenLong LI
b9068c1304
use volumes variable
fix leveldb mount bug
7 years ago
liwenlong05
02a1597642
fix leveldb mount bug
7 years ago
mtolman
e4b1e5c3c8
Fix for issue #635
Fix for the following issue on master: https://github.com/chrislusf/seaweedfs/issues/635
7 years ago
Yang Pengcheng
a2112e1be0
fix ldb dir deleting
7 years ago
Chris Lu
58344980e4
preallocate disk space during compaction also, add cleanup for failed compaction
7 years ago
Chris Lu
5047bdb4a2
skip bytes cache
8 years ago
Chris Lu
82c3ccc8dd
add btree for volume index
8 years ago
Chris Lu
bd97cbc523
skip busy reporting back to master
possible fix for https://github.com/chrislusf/seaweedfs/issues/499
8 years ago
Chris Lu
043b7a7c65
add locking to access l.volumes
8 years ago
brstgt
0656838fe5
Don't return actual file size from CheckVolumeDataIntegrity, it will be 0 if last needle is a tombstone, so it's not reliable anyway
8 years ago
brstgt
4fb5bb09b6
Remove obsolete property Volume.dataFileSize
8 years ago
brstgt
e074a54a20
Delete volumes online without restarting volume server
8 years ago
brstgt
492f93416d
Mount and unmount volumes online without restarting volume server
8 years ago
brstgt
18b3afc97a
Log volumeId if dataFileSize != actual data file size - improve log types
8 years ago
brstgt
0c898df430
Log volumeId if dataFileSize != actual data file size
8 years ago
Chris Lu
5b0f5f456e
volume server get notified of leader change
8 years ago
Chris Lu
59022b6fe0
fix byte counter on loading index file
fix https://github.com/chrislusf/seaweedfs/issues/441
8 years ago
Chris Lu
3065506b38
volume servers always connect to the master leader
8 years ago
Chris Lu
e46c3415f7
gRpc for master~volume heartbeat
8 years ago
Chris Lu
4beaaa0650
simplify the gzip optimization logic
fix https://github.com/chrislusf/seaweedfs/issues/436
8 years ago
Chris Lu
ed44f12f6d
support Fallocate on linux
8 years ago
Chris Lu
7b6837cbc2
move back the section
8 years ago
sparklxb
da9b672d1b
support additional header name-value pairs
8 years ago
sparklxb
86a7c56275
support additional header name-value pairs
8 years ago
Chris Lu
13e7069eb9
keep track of total data file size
8 years ago
lixianbin
d96d0a87cf
fix bug: upload big .gz file more than maxMB
8 years ago
Chris Lu
dcaf1796fe
add option to enable caching
8 years ago
Chris Lu
f54f530ada
adjust verbose logging
8 years ago
霍晓栋
b9f385bd66
refactor concurrent loading code logic
8 years ago
霍晓栋
096ffa9744
concurrent loading volume
8 years ago
Chris Lu
36f9633223
add locks for location.volumes
fix https://github.com/chrislusf/seaweedfs/issues/392
8 years ago
霍晓栋
7d73bbb073
comment UT case
8 years ago
霍晓栋
7d382ba5fe
comment UT case
8 years ago
霍晓栋
ce1f7ab662
makediff func with UT case
8 years ago
霍晓栋
ed848425c7
supplemental data between compacting and commit compacting
8 years ago
Chris Lu
dffad65f2f
fix compilation
8 years ago
Chris Lu
7e29218327
add a template for makeupDiff
8 years ago
Chris Lu
1bc041b46d
add a new way to manually compact corrupted volume
fix https://github.com/chrislusf/seaweedfs/issues/371
8 years ago
Chris Lu
01cbd5cb58
lock
fix https://github.com/chrislusf/seaweedfs/issues/367
8 years ago
Mike Tolman
0d331c1e3a
Revert "Changing needle_byte_cache so that it doesn't grow so big when larger files are added."
This reverts commit 87fee21ef5
.
9 years ago
Mike Tolman
87fee21ef5
Changing needle_byte_cache so that it doesn't grow so big when larger files are added.
9 years ago
霍晓栋
b9b3651a98
deleted index entry could not point to deleted needle
9 years ago
霍晓栋
09bd3d015d
deleted index entry could not point to deleted needle
9 years ago
霍晓栋
78678f4bcb
deleted needle does not need checksum verification
9 years ago
Chris Lu
d981eb282f
refactor volume_read_write.go out of volume.go
9 years ago
Chris Lu
576573711a
refactoring
9 years ago
Chris Lu
3d8df0f709
refactor volume_loading.go out of volume.go
9 years ago
Chris Lu
582d5d526c
refactoring
9 years ago
Chris Lu
3713119ab9
adjust verification logic.
1. log file name for easier debugging
2. remove unused version in readIndexEntryAtOffset()
3. if indexSize==0, the dataFile usually still has a super block of
size 8. So skipping this case.
9 years ago