Commit graph

53 commits

Author SHA1 Message Date
Aliaksandr Valialkin
ab8008d6d7 lib/{storage,mergeset}: remove empty directories on startup. Such directories can be left after unclean shutdown on NFS storage
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1142
2021-04-22 13:03:29 +03:00
Aliaksandr Valialkin
e03233f441 lib/fs: reduce the frequency of failed to remove directory ... due to NFS lock log warnings
Log `failed to remove directory ... due to NFS lock` warning only if the directory cannot be removed in one second.
2021-03-18 13:23:43 +02:00
Aliaksandr Valialkin
7f1302688f lib/fs: follow-up after f3a03c4164 2021-02-27 01:09:37 +02:00
Nikolay
d88fa5ebe4 Adds windows build (#1040)
* fixes windows compilation,
adds signal impl for windows,
adds free space usage for windows,
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/70
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1036

NOTE victoria metrics database still CANNOT work under windows system,
only vmagent is supported.
To completly port victoria metrics, you have to fix issues with separators,
parsing and posix file removall

* rollback separator

* Adds windows setInformation api,
it must behave like unix, need to test it.
changes procutil

* check for invlaid param

* Fixes posix delete semantic

* refactored a bit

* fixes openbsd build

* removed windows api call

* Fixes code after windows add

* Update lib/procutil/signal_windows.go

Co-authored-by: Aliaksandr Valialkin <valyala@gmail.com>
2021-02-27 01:06:22 +02:00
Aliaksandr Valialkin
c1b8729bd8 lib/fs: properly handle stale NFS file handle error during file deletion
This error can appear when -storageDataPath points to NFS volume and the given file has been already removed.
2021-02-26 23:24:46 +02:00
Aliaksandr Valialkin
148422bcba lib/storage: disable composite index usage when querying old data 2021-02-10 14:57:58 +02:00
Aliaksandr Valialkin
082eabf51e lib/fs: remove the code for tracking whether the given memory region is in page cache
This code didn't give performance gains under production workload, so let's remove it in order to simplify the code.
2021-02-09 16:51:11 +02:00
Aliaksandr Valialkin
588090765c lib/fs: properly initialize cleaner for pageCache bitmaps
Previously it wasnt working because the timer was fired only once
2021-01-27 00:40:39 +02:00
Aliaksandr Valialkin
ca8919e8e1 lib/storage: wait for pending transactions before closing and dropping the partition
This deflakes `make test-full-386` test
2020-12-25 11:46:47 +02:00
Aliaksandr Valialkin
f4fd917e4f lib/fs: replace fs.OpenReaderAt with fs.MustOpenReaderAt
All the callers for fs.OpenReaderAt expect that the file will be opened.
So it is better to log fatal error inside fs.MustOpenReaderAt instead of leaving this to the caller.
2020-11-23 09:57:30 +02:00
Aliaksandr Valialkin
97b836a6f4 lib/fs: fix GOOS=openbsd build by adding fadviseSequentialRead implementation.
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/785
2020-10-05 23:32:28 +03:00
Aliaksandr Valialkin
81cdf2fa14 lib/{fs,filestream}: small consistency-related updates after cc90a548b1 2020-09-29 00:43:20 +03:00
Nikolay Khramchikhin
658a05ef0f added openbsd implementations (#790)
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/785

removed fadvise for openbsd, added freespace implemenation for openbsd
2020-09-29 00:43:19 +03:00
Aliaksandr Valialkin
fd9f1463df lib/fs: use WARN instead of ERROR log level for the message when NFS diretory removal temporarily fails
this is expected condition, so it is better to use WARN log level for it
2020-08-09 12:07:35 +03:00
Aliaksandr Valialkin
f789e0fa44 lib/fs: export vm_nfs_pending_dirs_to_remove metric for monitoring the number of pending directories that couldn't be removed due to NFS lock 2020-08-06 15:31:50 +03:00
Aliaksandr Valialkin
77bb0e6595 lib/fs: clarify description for -fs.disableMmap command-line flag 2020-07-06 14:28:57 +03:00
Aliaksandr Valialkin
d962568e93 all: use %w instead of %s for wrapping errors in fmt.Errorf
This will simplify examining the returned errors such as httpserver.ErrorWithStatusCode .
See https://blog.golang.org/go1.13-errors for details.
2020-06-30 23:33:46 +03:00
Aliaksandr Valialkin
fd7a3d880e lib/fs: go fmt 2020-06-23 23:03:08 +03:00
Aliaksandr Valialkin
08edb90814 lib/fs: fall back to cgo copy for copying the last 4KB of mmaped data
This probably should fix https://github.com/VictoriaMetrics/VictoriaMetrics/issues/581
2020-06-23 22:55:56 +03:00
Aliaksandr Valialkin
521c657f8d lib/fs: an attempt to fix SIGBUS error by rounding mmap`ed region to multiple of 4KB pages
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/581
2020-06-23 13:40:20 +03:00
Aliaksandr Valialkin
ba1f764b29 lib/fs: optimize queries that read recent samples for big number of time series
Use standard copy() func instead of mmap-aware copy func for reading recently touched mmap-ed data.
This improves read performance by up to 4x.
2020-06-05 19:10:22 +03:00
Aliaksandr Valialkin
2358d9e41d lib/fs: add a benchmark for ReaderAt.MustReadAt 2020-06-05 19:10:21 +03:00
Aliaksandr Valialkin
3d0a0b3785 lib/fs: optimize MustGetFreeSpace performance by caching the results for up to 2 seconds 2020-06-04 13:14:04 +03:00
Aliaksandr Valialkin
c9ab6dc532 lib/fs: do not use mmap for 32-bit arches by default, since they cannot map files bigger than 4GB in RAM 2020-05-12 20:21:39 +03:00
Aliaksandr Valialkin
e933cbac16 lib/storage: postpone reading data from blocks during search
This eliminates the need for storing block data into temporary files on a single-node VictoriaMetrics
during heavy queries, which touch big number of time series over long time ranges.

This improves single-node VM performance on heavy queries by up to 2x.
2020-04-27 08:44:01 +03:00
Aliaksandr Valialkin
7a045125cc lib/fs: typo fix: read blocks bigger than 8KB via pread() call instead of using mmap 2020-02-25 18:04:06 +02:00
Aliaksandr Valialkin
7cde594696 all: do not clash flag description with back-quoted flag types
See https://golang.org/pkg/flag/#PrintDefaults for more details.
2020-02-04 15:56:01 +02:00
Aliaksandr Valialkin
9b25a2fb67 lib/fs: remove unused readerAt interface 2020-01-31 15:13:00 +02:00
Aliaksandr Valialkin
cb5c39ee70 lib/fs: optimize small reads for ReaderAt.MustReadAt by reading from memory-mapped space instead of reading from file descriptor
This should improve performance when reading many small blocks.
2020-01-30 15:16:16 +02:00
Aliaksandr Valialkin
da19fffa08 all: rename ReadAt* to MustReadAt* in order to dont clash with io.ReaderAt 2020-01-30 15:16:16 +02:00
Aliaksandr Valialkin
f22c9dbb0f lib/fs: typo fix in fadvise_unix.go 2019-12-24 21:00:04 +02:00
Aliaksandr Valialkin
6358cf3d47 app/vmselect/netstorage: move MustAdviseSequentialRead to lib/fs 2019-12-23 23:16:26 +02:00
Aliaksandr Valialkin
b9616c017f lib/{mergeset,storage}: remove transaction files only after the mentioned dirs are really removed
This should fix the issue on NFS when incompletely removed dirs may be left
after unclean shutdown (OOM, kill -9, hard reset, etc.), while the corresponding transaction
files are already removed.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/162
2019-12-02 21:34:37 +02:00
Aliaksandr Valialkin
87b39222be Revert "lib/fs: do not postpone directory removal on NFS error"
This reverts commit 21aeb02b46649ac9906cb37733f7b155a77a0db9.
2019-11-12 16:29:50 +02:00
Oleg Kovalov
74ba42d111 fix misspelled words (#229) 2019-11-12 00:18:24 +02:00
Aliaksandr Valialkin
5f52eb7653 lib/fs: do not postpone directory removal on NFS error
Continue trying to remove NFS directory on temporary errors for up to a minute.

The previous async removal process breaks in the following case during VictoriaMetrics start

- VictoriaMetrics opens index, finds incomplete merge transactions and starts replaying them.
- The transaction instructs removing old directories for parts, which were already merged into bigger part.
- VictoriaMetrics removes these directories, but their removal is delayed due to NFS errors.
- VictoriaMetrics scans partition directory after all the incomplete merge transactions are finished
  and finds directories, which should be removed, but weren't still removed due to NFS errors.
- VictoriaMetrics panics when it finds unexpected empty directory.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/162
2019-11-10 13:27:16 +02:00
Aliaksandr Valialkin
f581b2736a lib/fs: typo fix in comment to WriteFileAtomically 2019-10-29 11:31:34 +02:00
Aliaksandr Valialkin
2c654258ef lib/fs: add MustStopDirRemover for waiting until pending directories are removed on graceful shutdown
This patch is mainly required for laggy NFS. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/162
2019-09-05 11:17:17 +03:00
Aliaksandr Valialkin
82bfe818d0 lib/fs: try harder with directory removal on NFS in the event of temporary lock
Do not give up after 11 attempts of directory removal on laggy NFS.

Add `vm_nfs_dir_remove_failed_attempts_total` metric for counting the number of failed attempts
on directory removal.

Log failed attempts on directory removal after long sleep times.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/162
2019-09-04 12:24:41 +03:00
Aliaksandr Valialkin
604a4312f9 all: port to FreeBSD on GOARCH=amd64 2019-08-28 01:46:09 +03:00
Aliaksandr Valialkin
51263b1a45 lib/fs: add test for IsTemporaryFileName 2019-08-13 21:33:54 +03:00
Aliaksandr Valialkin
39f3f3a517 lib: move common code for creating flock.lock file into fs.CreateFlockFile 2019-08-13 01:46:20 +03:00
Aliaksandr Valialkin
73f866d874 lib/fs: atomically create file with the given contents on WriteFileAtomically
This should prevent from `transaction` and `metadata.json` files corruption
on unclean shutdown such as OOM, `kill -9`, power loss, etc.

Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/148
2019-08-12 15:02:04 +03:00
Aliaksandr Valialkin
41f512af1c all: add vm_data_size_bytes metrics for easy monitoring of on-disk data size and on-disk inverted index size 2019-07-04 19:43:04 +03:00
Aliaksandr Valialkin
b0b93e3d50 lib/fs: sync parent dir in MustRemoveAll only if it exists
The parent directory may be non-existing when the deleted directory
didn't exist before the MustRemoveAll call
2019-06-12 02:16:15 +03:00
Aliaksandr Valialkin
18d6f293f7 lib/fs: consolidate *RemoveAll* funcs into a single MustRemoveAll func
The func syncs parent dir in order to persist directory removal
in the event of power loss
2019-06-12 01:55:18 +03:00
Aliaksandr Valialkin
28d9904efc lib/fs: panic with fatal error when directories cannot be removed
Unremoved directories may lead to inconsistent data directory,
so VictoriaMetrics will fail to start next time.

So panic on the first error when trying to remove directory in order
to simplify recover process.
2019-06-12 01:20:10 +03:00
Aliaksandr Valialkin
d897bc3f08 lib/fs: attempt #2 to work around NFS issue with directory removal
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/61
2019-06-12 01:07:29 +03:00
Aliaksandr Valialkin
51e2e255a6 lib/fs: consistency renaming SyncPath -> MustSyncPath, since it doesnt return error 2019-06-11 23:13:45 +03:00
Aliaksandr Valialkin
3fa4c28f6b lib/fs: make sure the created directory remains visible in the fs in the event of power loss
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/63
2019-06-11 23:08:17 +03:00