107 commits

Author	SHA1	Message	Date
Aliaksandr Valialkin	343c69fc15	lib/{mergeset,storage}: pass compressLevel to blockStreamWriter.InitFromInmemoryPart ... This allows packing in-memory blocks with different compression levels depending on its contents. This may save memory usage.	2022-12-03 22:46:48 -08:00
Aliaksandr Valialkin	45299efe22	lib/{storage,mergeset}: consistency rename: `flushRaw{Rows,Items} -> flushPending{Rows,Items}	2022-12-03 22:17:46 -08:00
Aliaksandr Valialkin	e9636b4c69	lib/{mergeset,storage}: re-use the code for removing isInMerge flag at parts ... Move the common code into releasePartsToMerge() method and consistently use it throughout the code.	2022-12-02 18:52:37 -08:00
Aliaksandr Valialkin	c4265322f4	lib/fs: add canOverwrite arg to WriteFileAtomically when it is allowed to overwrite the file atomically if it already exists	2022-10-26 01:07:34 +03:00
Aliaksandr Valialkin	8e998aa1a1	lib/storage: add support for retention filters (aka multiple retentions for distinct sets of time series) ... Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/143 Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/289	2022-10-24 16:40:20 +03:00
Aliaksandr Valialkin	e2f0b76ebf	lib/storage: do not pass retentionMsecs and isReadOnly args explicitly - access them via Storage arg ... This makes code easier to read. This is a follow-up after d2d30581a0	2022-10-24 01:31:04 +03:00
Aliaksandr Valialkin	89a1108b1a	lib/storage: small code cleanups	2022-10-24 01:17:47 +03:00
Aliaksandr Valialkin	d2d30581a0	lib/storage: pass Storage to table and partition instead of getDeletedMetricIDs callback ... This improves code readability a bit.	2022-10-23 16:10:04 +03:00
Aliaksandr Valialkin	4128ad71e2	lib/storage: move common code to newRawRowsBlock() function	2022-10-21 14:46:55 +03:00
Aliaksandr Valialkin	b5674164c6	lib/storage: simplify code a bit after 3f5959c053	2022-10-21 14:39:27 +03:00
Aliaksandr Valialkin	fd7c86ae25	lib/{mergeset,storage}: simplify the code a bit after ae55ad8749	2022-10-21 14:33:03 +03:00
Aliaksandr Valialkin	3f5959c053	lib/storage: try generating initial parts from inmemory rows with identical sizes under high ingestion rate ... This should improve background merge rate under high load a bit	2022-10-20 23:28:24 +03:00
Aliaksandr Valialkin	150e99d403	lib/{mergeset,storage}: avoid unaligned 64-bit atomic operation panic on 32-bit platforms ... The panic has been introduced in 68f3a02589 While at it, add padding to shard structs in order to avoid false sharing on mordern CPUs This should improve scalability on systems with many CPU cores	2022-10-20 16:25:43 +03:00
Aliaksandr Valialkin	fb50730ba7	lib/storage: double the number of rawRows shards on multi-core systems ... This should increase data ingestion scalability on multi-core systems at the cost of slightly higher memory usage	2022-10-17 18:19:51 +03:00
Aliaksandr Valialkin	ae55ad8749	lib/{storage,mergeset}: do not hold per-shard lock in fast path when adding per-shard items to the flush list	2022-10-17 18:01:26 +03:00
Aliaksandr Valialkin	68e32b0764	lib/storage: atomically remove parts inside partitions ... Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3038	2022-09-13 16:17:38 +03:00
Aliaksandr Valialkin	340ada871d	lib/storage: atomically remove partitions, which went outside the configured retention ... Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3038	2022-09-13 16:17:37 +03:00
Aliaksandr Valialkin	9f94c295ab	all: use os.{Read|Write}File instead of ioutil.{Read|Write}File ... The ioutil.{Read|Write}File is deprecated since Go1.16 - see https://tip.golang.org/doc/go1.16#ioutil VictoriaMetrics needs at least Go1.18, so it is safe to remove ioutil usage from source code. This is a follow-up for 02ca2342ab	2022-08-21 23:52:35 +03:00
Roman Khavronenko	d59d829cdb	lib/storage: bump max merge concurrency for small parts to 15 (#2997) ... * lib/storage: bump max merge concurrency for small parts to 15 The change is based on the feedback from users on github. Thier examples show, that limit of 8 sometimes become a bottleneck. Users report that without limit concurrency can climb up to 15-20 merges at once. Signed-off-by: hagen1778 <roman@victoriametrics.com> * Update lib/storage/partition.go Signed-off-by: hagen1778 <roman@victoriametrics.com> Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>	2022-08-21 23:32:08 +03:00
Roman Khavronenko	a0e7432e42	lib/storage: prevent excessive loops when storage is in RO (#2962) ... * lib/storage: prevent excessive loops when storage is in RO Returning nil error when storage is in RO mode results into excessive loops and function calls which could result into CPU exhaustion. Returning an err instead will trigger delays in the for loop and save some resources. Signed-off-by: hagen1778 <roman@victoriametrics.com> * document the change Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>	2022-08-09 12:17:00 +03:00
Aliaksandr Valialkin	134751e43e	all: locate throttled loggers via logger.WithThrottler() only once and then use them ... This reduces the contention on logThrottlerRegistryMu mutex when logger.WithThrottler() is called frequently from concurrent goroutines.	2022-06-27 13:45:50 +03:00
Roman Khavronenko	1ee1e986da	lib/storage: limit max mergeConcurrency value for systems with high number of CPUs (#2673) ... Workers count for merges affects the max part size during merges. Such behaviour protects storage from running out of disk space for scenario when all workers are merging parts with the max size. This works very well for most cases. But for systems where high number of CPUs is allocated for vmstorage components this could significantly impact the max part size and result in more unmerged parts than expected. While checking multiple production highly loaded setups it was discovered that `max_over_time(vm_active_merges{type="storage/big}[1h]}"` rarely exceeds 2, and `max_over_time(vm_active_merges{type="storage/small}[1h]}"` rarely exceeds 4. The change in this commit limits the max value for concurrency accordingly. Signed-off-by: hagen1778 <roman@victoriametrics.com>	2022-06-07 14:55:09 +03:00
Aliaksandr Valialkin	ea06d2fd3c	lib/storage: stop background merge when storage enters read-only mode ... This should prevent from `no space left on device` errors when VictoriaMetrics under-estimates the additional disk space needed for background merge. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2603	2022-06-01 14:36:45 +03:00
Aliaksandr Valialkin	57143e9435	lib/storage: increase the number of rawRowsShard shards on systems with more than 4 CPU cores ... This should improve data ingestion scalability on systems with many CPU cores	2022-04-06 19:49:20 +03:00
Aliaksandr Valialkin	50cf74ce4b	lib/storage: reuse sync.WaitGroup objects ... This reduces GC load by up to 10% according to memory profiling	2022-04-06 13:34:04 +03:00
Aliaksandr Valialkin	59877d9f32	lib/{mergeset,storage}: tune compression levels for small blocks ... This should reduce CPU usage spent on compression	2022-02-25 15:33:40 +02:00
Aliaksandr Valialkin	145337792d	lib/{mergeset,storage}: properly limit cache sizes for indexdb ... Previously these caches could exceed limits set via `-memory.allowedPercent` and/or `-memory.allowedBytes`, since limits were set independently per each data part. If the number of data parts was big, then limits could be exceeded, which could result to out of memory errors. Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2007	2022-01-20 18:37:17 +02:00
Aliaksandr Valialkin	ce333f28d8	all: use logger.WithThrottler() where appropriate	2021-12-21 17:03:25 +02:00
Aliaksandr Valialkin	8a7f08ded3	lib/storage: properly update per-part min_dedup_interval file contents after merge ... Previously 0s was always written even if -dedup.minScrapeInterval was set to non-zero value This is a follow-up for 4ff647137a	2021-12-17 20:13:24 +02:00
Aliaksandr Valialkin	4ff647137a	lib/storage: deduplicate samples more thoroughly ... Previously some duplicate samples may be left on disk for time series with high churn rate. This may result in higher disk space usage.	2021-12-15 15:59:58 +02:00
Aliaksandr Valialkin	7275ebf91a	app/vmstorage: export vm_cache_size_max_bytes metrics for determining capacity of various caches ... The vm_cache_size_max_bytes metric can be used for determining caches which reach their capacity via the following query: vm_cache_size_bytes / vm_cache_size_max_bytes > 0.9	2021-12-02 10:30:43 +02:00
Aliaksandr Valialkin	2fb5a6ca78	lib/storage: do not take into account -storage.minFreeDiskSpaceBytes during background merges	2021-12-01 11:02:36 +02:00
Aliaksandr Valialkin	d666755159	lib/storage: take into account -storage.minFreeDiskSpaceBytes when performing big merges ... Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/269	2021-11-30 12:56:35 +02:00
Aliaksandr Valialkin	ffc0ab1774	lib/{mergeset,storage}: improve the detection of the needed free space for background merge ... This should prevent from possible out of disk space crashes during big merges. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1560	2021-08-25 09:35:44 +03:00
Aliaksandr Valialkin	c2deee9911	lib/storage: yet another attempt to properly determine disk space shortage, which prevents from optimal merges ... Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1373	2021-07-27 12:04:50 +03:00
Aliaksandr Valialkin	9a83e9018d	lib/storage: properly detect free disk space shortage during data merge ... Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1373	2021-07-02 17:40:54 +03:00
Aliaksandr Valialkin	dcbc22552f	lib/storage: fix infinite loop introduced in aa9b56a046	2021-06-17 14:28:10 +03:00
Aliaksandr Valialkin	aa9b56a046	lib/{mergeset,storage}: reduce the number of fsync calls on data ingestion path on systems with many cpu cores ... VictoriaMetrics maintains a buffer per CPU core for the ingested data. These buffers are flushed to disk every second. These buffers are flushed to disk in parallel starting from the commit 56b6b893ce . This resulted in increased write disk IO usage on systems with many cpu cores as described at https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338#issuecomment-863046999 . This commit merges the per-CPU buffers into bigger in-memory buffers before flushing them to disk. This should reduce the rate of fsync syscalls and, consequently, the write disk IO on systems with many CPU cores. This should help https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1338 See also https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1244	2021-06-17 13:52:08 +03:00
Aliaksandr Valialkin	69b1482bdb	lib/storage: consistency renaming: getMaxRawRowsPerPartition -> getMaxRawRowsPerShard	2021-06-11 10:57:23 +03:00
Aliaksandr Valialkin	044ab46824	lib/storage: reduce the amounts of memory which can be occupied by rawRow items during data ingestion on a system with many CPU cores	2021-06-11 10:57:23 +03:00
Aliaksandr Valialkin	a4ff4b8e65	lib/storage: allow filling all the rows up to their capacity in rawRowsShard.addRows ... This should reduce memory usage a bit on data ingestion path	2021-05-24 15:22:59 +03:00
Aliaksandr Valialkin	ec79abc382	lib/{mergeset,storage}: reduce the number of IFNO log messages like merged ... items across ... blocks in ... seconds ... Log these messages if the merge takes more than 30 seconds instead of 10 seconds.	2021-05-23 14:03:21 +03:00
Nikolay	477369b62f	adds stalePartsRemover (#1261) ... for new created partitions	2021-05-03 11:34:00 +03:00
Aliaksandr Valialkin	87179c6839	lib/{storage,mergeset}: fix unaligned 64-bit atomic operation panic for 32-bit architectures ... The panic has been introduced in 56b6b893ce	2021-04-27 16:41:32 +03:00
Aliaksandr Valialkin	56b6b893ce	lib/mergeset: split rows ingestion among multiple shards ... This improves rows ingestion on systems with many CPU cores by reducing lock contention. Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1244 Thanks to @waldoweng for the original idea and draft implementation at https://github.com/VictoriaMetrics/VictoriaMetrics/pull/1243	2021-04-27 15:36:34 +03:00
Aliaksandr Valialkin	f89c1f7f49	lib/storage: typo fix in info message when deleting the part outside the configured retention ... Previously the message was displaying incorrect retention time	2021-04-27 13:32:46 +03:00
Aliaksandr Valialkin	bbebdf9ba1	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 13:02:44 +03:00
Aliaksandr Valialkin	48656dcc38	lib/{mergeset,storage}: allow merging smaller number of small parts ... While this may increase CPU and disk IO usage needed for background merge, this also recudes CPU usage during queries in production. This is because such queries tend to read recently added data and it is better to have lower number of parts for such data in order to reduce CPU usage. This partially reverts ebf8da3730	2021-02-21 21:28:36 +02:00
Aliaksandr Valialkin	cb96a1865b	app/vmstorage: export missing vm_cache_size_bytes metrics for indexdb and data caches	2021-02-09 00:47:00 +02:00
Aliaksandr Valialkin	9dcb18e03d	app/vmstorage: disable final merge by default, since it may result in high disk IO and CPU usage without measurable benefits such as increased query performance and reduced disk space usage	2021-01-08 00:16:05 +02:00