Commit graph

649 commits

Author SHA1 Message Date
Chris Lu 2cca07b44c cloud tier to non-AWS s3 gateways 2021-02-28 18:59:09 -08:00
Chris Lu 30b30b8fe0 volume.tier.move: passing non-empty disk type 2021-02-22 01:59:03 -08:00
Chris Lu 03c643aa6b fix test
fileSize need to be divided by 8
2021-02-20 21:19:21 -08:00
Chris Lu 7635f6b9fa disk file avoid file.Stat() 2021-02-20 20:06:06 -08:00
Chris Lu a2383b3b12 Revert "Revert "avoid file.Stat()""
This reverts commit 7ef4c24f18.
2021-02-20 12:42:09 -08:00
Chris Lu 7ef4c24f18 Revert "avoid file.Stat()"
This reverts commit 98c93ca465.
2021-02-20 12:39:33 -08:00
Chris Lu 1ad3200094 skip seek() when index file writes 2021-02-20 12:39:25 -08:00
Chris Lu 98c93ca465 avoid file.Stat()
avoid one Syscall, but did not help on performance though
2021-02-20 09:44:17 -08:00
Chris Lu a1210d1e8d Revert "volume: avoid file.stat(), file.seek() if possible during writes"
This reverts commit c78409a598.
2021-02-19 03:56:27 -08:00
Chris Lu c78409a598 volume: avoid file.stat(), file.seek() if possible during writes 2021-02-18 23:22:54 -08:00
Chris Lu 73958e357d add descriptive error if no free volumes 2021-02-18 19:10:20 -08:00
bingoohuang 352ac2f271 Merge remote-tracking branch 'origin/master' 2021-02-18 14:05:51 +08:00
bingoohuang eab6e31d34 use backticks instead of double quotes to avoid escaped additionally in regex 2021-02-18 14:05:28 +08:00
Chris Lu 3575d41009 go fmt 2021-02-17 20:57:08 -08:00
Chris Lu 7403cd43c5 only use "" for hdd to avoid two values for the same thing 2021-02-16 10:50:42 -08:00
Chris Lu ebf320ec22 avoid divided by zero 2021-02-16 05:59:24 -08:00
Chris Lu 3fe628f04e use hdd instead of empty string 2021-02-16 03:03:00 -08:00
Chris Lu f8446b42ab this can compile now!!! 2021-02-16 02:47:02 -08:00
Chris Lu 7ce647f27e support customizable disk type 2021-02-13 15:42:42 -08:00
Chris Lu 4bd8a692d8 disk type can be generic tags 2021-02-13 13:50:14 -08:00
Chris Lu 885ca34748 volume: fail fast if idx files are missing
fix https://github.com/chrislusf/seaweedfs/issues/1796
2021-02-11 00:44:40 -08:00
Chris Lu 821c46edf1 Merge branch 'master' into support_ssd_volume 2021-02-09 11:37:07 -08:00
bingoohuang 9e06ef66b9 fix renaming for 5bytes 2021-02-09 09:44:28 +08:00
Chris Lu a833021132 fix refactoring left over 2021-02-07 23:03:03 -08:00
bingoohuang 7256902fb0 fix typo offset.ToAcutalOffset to offset.ToActualOffset 2021-02-07 12:11:51 +08:00
bingoohuang 94ea3bd3a5 renaming NeedleMapType to NeedleMapKind 2021-02-07 09:00:03 +08:00
Chris Lu cfb9342a15 avoid concurrent map updates to viper 2021-01-12 02:28:13 -08:00
Chris Lu 039ad101b1 manifest file also need to detect compression
fix https://github.com/chrislusf/seaweedfs/issues/1724#issuecomment-753585046
2021-01-03 01:42:21 -08:00
Chris Lu 4f31c1bb94 go fmt 2020-12-22 02:34:08 -08:00
Chris Lu 9abf016af1 Merge branch 'master' into support_ssd_volume 2020-12-17 22:30:19 -08:00
Chris Lu 738c4fd203 fix logging when error
address https://github.com/chrislusf/seaweedfs/pull/1685#issuecomment-747845601
2020-12-17 19:37:46 -08:00
Chris Lu daa8157fc2 Merge branch 'master' into support_ssd_volume 2020-12-17 13:05:20 -08:00
Chris Lu e2076201d7 volume: avoid reprocessing the same volume
fix https://github.com/chrislusf/seaweedfs/issues/1682
2020-12-17 13:03:39 -08:00
Chris Lu 986cbdf7d9 Revert "Merge pull request #1683 from qieqieplus/master"
This reverts commit 8cb67952db, reversing
changes made to 200e56215a.
2020-12-17 12:46:42 -08:00
Chris Lu f56e6d231e Revert "Merge pull request #1683 from qieqieplus/master"
This reverts commit 8cb67952db, reversing
changes made to 200e56215a.
2020-12-17 12:46:20 -08:00
Chris Lu 1bf22c0b5b go fmt 2020-12-16 09:14:05 -08:00
Chris Lu 1d88865869 passing disk type along 2020-12-16 09:10:14 -08:00
Chris Lu 23014b6810 Merge branch 'master' into support_ssd_volume 2020-12-16 08:26:51 -08:00
chenqieqie 45f902a9b7 smaller critical section 2020-12-16 20:23:51 +08:00
chenqieqie 4e58a4f24e fix race condition when loading volumes concurrently 2020-12-16 18:49:10 +08:00
Chris Lu 2e8dba571b adjust volume server UI 2020-12-14 00:51:57 -08:00
Chris Lu 933f2f4cfd find location with matching disk type 2020-12-13 23:39:00 -08:00
Chris Lu 94525aa0fd allocate volume by disk type 2020-12-13 23:08:21 -08:00
Chris Lu f6a419c26c disk type configurable for each folder 2020-12-13 22:49:56 -08:00
Chris Lu 0d2ec832e2 rename from volumeType to diskType 2020-12-13 11:59:32 -08:00
Chris Lu d156c74ec0 volume server set volume type and heartbeat to the master 2020-12-13 03:11:24 -08:00
Chris Lu e9cd798bd3 adding volume type 2020-12-13 00:58:58 -08:00
Chris Lu ae655033ac adjust logging 2020-12-11 16:57:53 -08:00
Chris Lu d986c7196d use append time instead of filer's own modification time
fix https://github.com/chrislusf/seaweedfs/issues/1669
2020-12-11 16:55:18 -08:00
Chris Lu d171d9f988 volume: ensure the volume dat and idx files are always closed
fix https://github.com/chrislusf/seaweedfs/issues/1646
2020-12-04 21:50:26 -08:00
Chris Lu cc839f935d better unit detection 2020-12-03 19:23:59 -08:00
Chris Lu b219ccfe68 Update volume_ttl_test.go 2020-12-03 12:55:44 -08:00
Chris Lu a83b8e7b01 ttl fix: 3y was converted into 96m 2020-12-03 10:53:29 -08:00
Chris Lu f3bb645018 file open error 2020-12-01 23:37:49 -08:00
Chris Lu dc0bc48257 return file open error 2020-12-01 23:36:49 -08:00
Chris Lu 96c48bc8a8 fix test 2020-11-28 03:03:43 -08:00
Chris Lu 2c913dde04 volume: detect and drop volumes with disk IO error
from Jethro in slack:
is it possible to make the assign request a bit smarter? Currently I’m in the state that a disk failed but all assign request are being send to this volume. It would be cool if the master sees this and stopped using this volume.
e=HTTP(http://x:8089/913,045a782b63176edf) not 200 but 500 Internal Server Error
Body={"size":740167,"error":"failed to write to local disk: write /mnt/v9/913.dat: input/output error","eTag":"ee4381e202212ff3aee647704c036689"}
e=HTTP(http://x:8089/913,045a782c90240077) not 200 but 500 Internal Server Error
Body={"size":792779,"error":"failed to write to local disk: write /mnt/v9/913.dat: input/output error","eTag":"c43463ccc11eb6eb2fc306f407a6a953"}
e=HTTP(http://x:8089/913,045a782e6b7901ea) not 200 but 500 Internal Server Error
Body={"size":3962392,"error":"failed to write to local disk: write /mnt/v9/913.dat: input/output error","eTag":"04c91198e9b276c81f11dbf189af5d28"}
2020-11-28 00:09:29 -08:00
Chris Lu 9ac4935f22 read from volume index file directly instead of open a separate file
fix https://github.com/chrislusf/seaweedfs/issues/1640

read from volume index file directly instead of open a separate file,
to ensure reading latest index entries.
2020-11-27 16:18:48 -08:00
Chris Lu 6d30b21b10 volume: add "-dir.idx" option for separate index storage
fix https://github.com/chrislusf/seaweedfs/issues/1265
2020-11-27 03:17:10 -08:00
Chris Lu 3c229eb677 refactor 2020-11-26 17:21:55 -08:00
Chris Lu c0fb4a3766 add comments 2020-11-26 15:22:42 -08:00
Chris Lu 6783ae5f8f rename 2020-11-26 15:19:43 -08:00
Chris Lu 97f706f35b just refactoring 2020-11-26 15:08:05 -08:00
Chris Lu b11449f955 pass in dir.idx parameter to DiskLocation 2020-11-26 14:59:03 -08:00
Chris Lu 0da7ecfd29 go fmt 2020-11-26 11:22:30 -08:00
Chris Lu 92f906b6fc remove zstd
fix https://github.com/chrislusf/seaweedfs/issues/1629
2020-11-21 13:06:45 -08:00
Chris Lu 1dd3a6ac36 avoid verbose logs 2020-11-09 15:56:11 -08:00
Chris Lu 0dafcf1f5a volume: detect max volume count changes based on disk usage
fix https://github.com/chrislusf/seaweedfs/issues/1594
2020-11-03 14:43:17 -08:00
Chris Lu de86945aeb go fmt 2020-10-31 16:45:38 -07:00
Chris Lu 9708df47d1 ensure offset in the index file is positive 2020-10-31 15:50:01 -07:00
Konstantin Lebedev 46303c36bf When the volume server is stopped, the master server immediately sees the deletion of volumes 2020-10-31 02:48:25 +05:00
Chris Lu 6560ac6466 volume loading: trim out unreachable idx file content
fix https://github.com/chrislusf/seaweedfs/issues/1583
2020-10-29 22:25:23 -07:00
Konstantin Lebedev 1ad1b8c4f6 collection Volume ReadOnly Count with detailed status 2020-10-28 17:38:26 +05:00
Chris Lu 744c5594da volume: automatically trim out unreachable entries 2020-10-28 01:14:39 -07:00
Chris Lu 3e925faddd trim out name extension 2020-10-28 01:14:05 -07:00
shibinbin a67eb1afa1 fix: restart volumeserver sometime occurred out of memory error 2020-10-28 12:03:06 +08:00
Chris Lu 53c3aad875 volume: add a note file to avoid incomplete volume files
fix https://github.com/chrislusf/seaweedfs/issues/1567
2020-10-27 15:56:49 -07:00
Chris Lu 06c15ab35c volume: add special handling for .dat larger than 32GB 2020-10-27 13:11:56 -07:00
Konstantin Lebedev 2fb1fce8a0 avoid old values when missing read only flags 2020-10-25 16:21:40 +05:00
Chris Lu 9104cfa744 reduce locks 2020-10-24 19:40:35 -07:00
Chris Lu e9d40b80b2 less verbose logs 2020-10-24 09:42:54 -07:00
Chris Lu 575d7952a1 add available resource stats
fix https://github.com/chrislusf/seaweedfs/issues/1555
2020-10-22 09:13:47 -07:00
Chris Lu 4ff2ceee33 UI fix on rendering EC volumes
addressing UI problem with https://github.com/chrislusf/seaweedfs/issues/1551
2020-10-21 22:05:58 -07:00
Konstantin Lebedev dc2e13092d add number of read only volumes metric 2020-10-15 15:48:40 +05:00
Chris Lu bd8b9b0c2e some changes to prepare fix after failed reads 2020-10-08 18:33:06 -07:00
Chris Lu d522df0ecd adjust log level 2020-10-04 01:35:32 -07:00
Chris Lu a1c01d716b volume: avoid deadlock when deleting volumes
fix https://github.com/chrislusf/seaweedfs/issues/1501
2020-10-01 07:10:03 -07:00
Chris Lu 103fafe00b weed export: print out [start,stop) content range 2020-09-14 22:57:23 -07:00
Chris Lu f2723c1bc8 do not idx file format
revert c9ab8d05fa
2020-09-12 12:42:36 -07:00
Chris Lu 446e476a11 go fmt 2020-09-12 04:08:03 -07:00
Chris Lu ea26a98753 volume: validate volume correctness if last entry is a deletion 2020-09-12 04:07:04 -07:00
Chris Lu 9d4bdfcfdf fix volume integrity checking 2020-09-11 11:34:10 -07:00
Chris Lu 401ccf1509 fix test 2020-09-10 15:13:23 -07:00
Chris Lu c9ab8d05fa fixes for reading deleted fid 2020-09-10 14:42:52 -07:00
James Hartig 8e54e34576 volume: Don't unmount before deleting volume in copy
If we unmount first and then delete, the delete fails because the volume
was unmounted. Delete ends up doing the same thing as the unmount anyways.
2020-09-01 22:00:07 -04:00
Chris Lu 38e06d783d volume: check disk space before compaction
fix https://github.com/chrislusf/seaweedfs/issues/1440
2020-08-31 18:10:53 -07:00
Chris Lu dbb10e0f0b testing with more than 30days 2020-08-31 17:59:29 -07:00
Chris Lu 99ecf63276 go fmt 2020-08-29 22:28:33 -07:00
Chris Lu ab759f0ec2 erasure coding: fix EC error if multiple disks are configured in one volume server 2020-08-26 09:16:58 -07:00
Chris Lu d1cf39f180 fix logging 2020-08-26 08:55:15 -07:00