- Maintain a separate worker pool per each part type (in-memory, file, big and small).
Previously a shared pool was used for merging all the part types.
A single merge worker could merge parts with mixed types at once. For example,
it could merge simultaneously an in-memory part plus a big file part.
Such a merge could take hours for big file part. During the duration of this merge
the in-memory part was pinned in memory and couldn't be persisted to disk
under the configured -inmemoryDataFlushInterval .
Another common issue, which could happen when parts with mixed types are merged,
is uncontrolled growth of in-memory parts or small parts when all the merge workers
were busy with merging big files. Such growth could lead to significant performance
degradataion for queries, since every query needs to check ever growing list of parts.
This could also slow down the registration of new time series, since VictoriaMetrics
searches for the internal series_id in the indexdb for every new time series.
The third issue is graceful shutdown duration, which could be very long when a background
merge is running on in-memory parts plus big file parts. This merge couldn't be interrupted,
since it merges in-memory parts.
A separate pool of merge workers per every part type elegantly resolves both issues:
- In-memory parts are merged to file-based parts in a timely manner, since the maximum
size of in-memory parts is limited.
- Long-running merges for big parts do not block merges for in-memory parts and small parts.
- Graceful shutdown duration is now limited by the time needed for flushing in-memory parts to files.
Merging for file parts is instantly canceled on graceful shutdown now.
- Deprecate -smallMergeConcurrency command-line flag, since the new background merge algorithm
should automatically self-tune according to the number of available CPU cores.
- Deprecate -finalMergeDelay command-line flag, since it wasn't working correctly.
It is better to run forced merge when needed - https://docs.victoriametrics.com/#forced-merge
- Tune the number of shards for pending rows and items before the data goes to in-memory parts
and becomes visible for search. This improves the maximum data ingestion rate and the maximum rate
for registration of new time series. This should reduce the duration of data ingestion slowdown
in VictoriaMetrics cluster on e.g. re-routing events, when some of vmstorage nodes become temporarily
unavailable.
- Prevent from possible "sync: WaitGroup misuse" panic on graceful shutdown.
This is a follow-up for fa566c68a6 .
Thanks @misutoth to for the inspiration at https://github.com/VictoriaMetrics/VictoriaMetrics/pull/5212
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/5190
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3790
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3551
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3337
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3425
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3647
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3641
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/648
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/291
This allows reducing the indexdb/tagFiltersToMetricIDs cache size by 8 on average.
The cache size can be checked via vm_cache_size_bytes{type="indexdb/tagFiltersToMetricIDs"} metric exposed at /metrics page.
Previously, it was not possible to determine which tenant sends metrics with excessive amount of labels of label values.
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
This should smooth CPU and RAM usage spikes related to these periodic tasks,
by reducing the probability that multiple concurrent periodic tasks are performed at the same time.
The dateMetricIDCache puts recently registered (date, metricID) entries into mutable cache protected by the mutex.
The dateMetricIDCache.Has() checks for the entry in the mutable cache when it isn't found in the immutable cache.
Access to the mutable cache is protected by the mutex. This means this access is slow on systems with many CPU cores.
The mutabe cache was merged into immutable cache every 10 seconds in order to avoid slow access to mutable cache.
This means that ingestion of new time series to VictoriaMetrics could result in significant slowdown for up to 10 seconds
because of bottleneck at the mutex.
Fix this by merging the mutable cache into immutable cache after len(cacheItems) / 2
cache hits under the mutex, e.g. when the entry is found in the mutable cache.
This should automatically adjust intervals between merges depending on the addition rate
for new time series (aka churn rate):
- The interval will be much smaller than 10 seconds under high churn rate.
This should reduce the mutex contention for mutable cache.
- The interval will be bigger than 10 seconds under low churn rate.
This should reduce the uneeded work on merging of mutable cache into immutable cache.
- Clarify the bugfix description at docs/CHANGELOG.md
- Simplify the code by accessing prefetchedMetricIDs struct under the lock
instead of using lockless access to immutable struct.
This shouldn't worsen code scalability too much on busy systems with many CPU cores,
since the code executed under the lock is quite small and fast.
This allows removing cloning of prefetchedMetricIDs struct every time
new metric names are pre-fetched. This should reduce load on Go GC,
since the cloning of uin64set.Set struct allocates many new objects.
Before, this cache was limited only by size.
Cache invalidation by time happens with jitter to prevent thundering herd problem.
Signed-off-by: hagen1778 <roman@victoriametrics.com>
* app/vmselect: limit the number of parallel workers by 32
The change should improve performance and memory usage during query processing
on machines with big number of CPU cores. The number of parallel workers for
query processing is controlled via `-search.maxWorkersPerQuery` command-line flag.
By default, the number of workers is limited by the number of available CPU cores,
but not more than 32. The limit can be increased via `-search.maxWorkersPerQuery`.
Signed-off-by: hagen1778 <roman@victoriametrics.com>
* wip
- The `-search.maxWorkersPerQuery` command-line flag doesn't limit resource usage,
so move it from the `resource usage limits` to `troubleshooting` chapter at docs/Single-server-VictoriaMetrics.md
- Make more clear the description for the `-search.maxWorkersPerQuery` command-line flag
- Add the description of `-search.maxWorkersPerQuery` to docs/Cluster-VictoriaMetrics.md
- Limit the maximum value, which can be passed to `-search.maxWorkersPerQuery`, to GOMAXPROCS,
because bigger values may worsen query performance and increase CPU usage
- Improve the the description of the change at docs/CHANGELOG.md. Mark it as FEATURE instead of BUGFIX,
since it is closer to a feature than to a bugfix.
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/5087
---------
Signed-off-by: hagen1778 <roman@victoriametrics.com>
Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
This should reduce tail latency during data ingestion.
This shouldn't slow down data ingestion in the worst case, since assisted merges are spread among
distinct addRows/addItems calls after this change.
* lib/storage/partition: add check to ensure parts exist on disk
If part exists in parts.json but is missing on disk there will be a misleading error similar to "unexpected number of substrings in the part name".
This change forces verification of part existence and throws a correct error in case it is missing on disk.
Such issue can be result of https://github.com/VictoriaMetrics/VictoriaMetrics/issues/5005 or disk corruption.
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
* lib/storage/partition: use filepath.Join instead of string concatenation
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
* lib/storage/partition: add action points for error message
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
* all: add a check for missing part in lib/mergeset and lib/logstorage
---------
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
* Introduce flagutil.Duration
To avoid conversion bugs
* Fix tests
* Clarify documentation re. month=31 days
* Add fasttime.UnixTime() to obtain time.Time
The goal is to refactor out the last usage of `.Msecs`.
* Use fasttime for time.Now()
* wip
- Remove fasttime.UnixTime(), since it doesn't improve code readability and maintainability
- Run `make docs-sync` for syncing changes from README.md to docs/ folder
- Make lib/flagutil.Duration.Msec private
- Rename msecsPerMonth const to msecsPer31Days in order to be consistent with retention31Days
---------
Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
* lib/storage: pre-create timeseries before indexDB rotation
during an hour before indexDB rotation start creating records at the next indexDB
it must improve performance during switch for the next indexDB and remove ingestion issues.
Since there is no need for creation new index records for timeseries already ingested into current indexDB
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/4563
* lib/storage: further work on indexdb rotation optimization
- Document the change at docs/CHAGNELOG.md
- Move back various caches from indexDB to Storage. This makes the change less intrusive.
The dateMetricIDCache now takes into account indexDB generation, so it stores (date, metricID)
entries for both the current and the next indexDB.
- Consolidate the code responsible for idbNext pre-filling into prefillNextIndexDB() function.
This improves code readability and maintainability a bit.
- Rewrite and simplify the code responsible for calculating the next retention timestamp.
Add various tests for corner cases of this code.
- Remove indexdb pre-filling from RegisterMetricNames() function, since this function is rarely called.
It is OK to add indexdb entries on demand in this function. This simplifies the code.
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1401
* docs/CHANGELOG.md: refer to https://github.com/VictoriaMetrics/VictoriaMetrics/issues/4563
---------
Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
This eliminates the need in .(*T) casting for results obtained from Load()
Leave atomic.Value for map, since atomic.Pointer[map[...]...] makes double pointer to map,
because map is already a pointer type.
- Sort MetricName tags only once before the benchmark loop.
- Obtain indexSearch per each benchmark loop in order to give a chance for background merge
for the recently created parts
Previously all the newly ingested time series were registered in global `MetricName -> TSID` index.
This index was used during data ingestion for locating the TSID (internal series id)
for the given canonical metric name (the canonical metric name consists of metric name plus all its labels sorted by label names).
The `MetricName -> TSID` index is stored on disk in order to make sure that the data
isn't lost on VictoriaMetrics restart or unclean shutdown.
The lookup in this index is relatively slow, since VictoriaMetrics needs to read the corresponding
data block from disk, unpack it, put the unpacked block into `indexdb/dataBlocks` cache,
and then search for the given `MetricName -> TSID` entry there. So VictoriaMetrics
uses in-memory cache for speeding up the lookup for active time series.
This cache is named `storage/tsid`. If this cache capacity is enough for all the currently ingested
active time series, then VictoriaMetrics works fast, since it doesn't need to read the data from disk.
VictoriaMetrics starts reading data from `MetricName -> TSID` on-disk index in the following cases:
- If `storage/tsid` cache capacity isn't enough for active time series.
Then just increase available memory for VictoriaMetrics or reduce the number of active time series
ingested into VictoriaMetrics.
- If new time series is ingested into VictoriaMetrics. In this case it cannot find
the needed entry in the `storage/tsid` cache, so it needs to consult on-disk `MetricName -> TSID` index,
since it doesn't know that the index has no the corresponding entry too.
This is a typical event under high churn rate, when old time series are constantly substituted
with new time series.
Reading the data from `MetricName -> TSID` index is slow, so inserts, which lead to reading this index,
are counted as slow inserts, and they can be monitored via `vm_slow_row_inserts_total` metric exposed by VictoriaMetrics.
Prior to this commit the `MetricName -> TSID` index was global, e.g. it contained entries sorted by `MetricName`
for all the time series ever ingested into VictoriaMetrics during the configured -retentionPeriod.
This index can become very large under high churn rate and long retention. VictoriaMetrics
caches data from this index in `indexdb/dataBlocks` in-memory cache for speeding up index lookups.
The `indexdb/dataBlocks` cache may occupy significant share of available memory for storing
recently accessed blocks at `MetricName -> TSID` index when searching for newly ingested time series.
This commit switches from global `MetricName -> TSID` index to per-day index. This allows significantly
reducing the amounts of data, which needs to be cached in `indexdb/dataBlocks`, since now VictoriaMetrics
consults only the index for the current day when new time series is ingested into it.
The downside of this change is increased indexdb size on disk for workloads without high churn rate,
e.g. with static time series, which do no change over time, since now VictoriaMetrics needs to store
identical `MetricName -> TSID` entries for static time series for every day.
This change removes an optimization for reducing CPU and disk IO spikes at indexdb rotation,
since it didn't work correctly - see https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1401 .
At the same time the change fixes the issue, which could result in lost access to time series,
which stop receving new samples during the first hour after indexdb rotation - see https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2698
The issue with the increased CPU and disk IO usage during indexdb rotation will be addressed
in a separate commit according to https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1401#issuecomment-1553488685
This is a follow-up for 1f28b46ae9
The number of parts in the snapshot partition may be zero if concurrent goroutine just
started creating new partition, but didn't put data into it yet when the current
goroutine made a snapshot.
* lib/storage: creates parts.json on start-up if it not exists.
It fixes migrations from versions below v1.90.0.
Previously parts.json was created only after successful merge.
But if merge was interruped for some reason (OOM or shutdown), parts.json wasn't created and partitions left after interruped merge weren't properly deleted.
Since VM cannot check if it must be removed or not.
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/4336
* Apply suggestions from code review
Co-authored-by: Roman Khavronenko <roman@victoriametrics.com>
* Update lib/storage/partition.go
Co-authored-by: Roman Khavronenko <roman@victoriametrics.com>
---------
Co-authored-by: Roman Khavronenko <roman@victoriametrics.com>
This reverts the following commits:
- e0e16a2d36
- 2ce02a7fe6
The reason for revert: the updated logic breaks assumptions made
when fixing https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2698 .
For example, if a time series stop receiving new samples during the first
day after the indexdb rotation, there are chances that the time series
won't be registered in the new indexdb. This is OK until the next indexdb
rotation, since the time series is registered in the previous indexdb,
so it can be found during queries. But the time series will become invisible
for search after the next indexdb rotation, while its data is still there.
There is also incompletely solved issue with the increased CPU and disk IO resource
usage just after the indexdb rotation. There was an attempt to fix it, but it didn't fix
it in full, while introducing the issue mentioned above. See https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1401
TODO: to find out the solution, which simultaneously solves the following issues:
- increased memory usage for setups high churn rate and long retention (e.g. what the reverted commit does)
- increased CPU and disk IO usage during indexdb rotation ( https://github.com/VictoriaMetrics/VictoriaMetrics/issues/1401 )
- https://github.com/VictoriaMetrics/VictoriaMetrics/issues/2698
- Document the change at docs/CHANGELOG.md
- Clarify comments for non-trivial code touched by the commit
- Improve the logic behind maybeCreateIndexes():
- Correctly create per-day indexes if the indexdb rotation is performed during
the first hour or the last hour of the day by UTC.
Previously there was a possibility of missing index entries on that day.
- Increase the duration for creating new indexes in the current indexdb for up to 22 hours
after indexdb rotation. This should reduce the increased resource usage
after indexdb rotation.
It is safe to postpone index creation for the current day until the last hour
of the current day after indexdb rotation by UTC, since the corresponding (date, ...)
entries exist in the previous indexdb.
- Search for TSID by (date, MetricName) in both the current and the previous indexdb.
Previously the search was performed only in the current indexdb. This could lead
to excess creation of per-day indexes for the current day just after indexdb rotation.
- Search for (date, metricID) entries in both the current and the previous indexdb.
Previously the search was performed only in the current indexdb. This could lead
to excess creation of per-day indexes for the current day just after indexdb rotation.
The new index substitutes global MetricName=>TSID index
used for locating TSIDs on ingestion path.
For installations with high ingestion and churn rate, global
MetricName=>TSID index can grow enormously making
index lookups too expensive. This also results into bigger
than expected cache growth for indexdb blocks.
New per-day index supposed to be much smaller and more efficient.
This should improve ingestion speed and reliability during
re-routings in cluster.
The negative outcome could be occupied disk size, since
per-day index is more expensive comparing to global index.
Signed-off-by: hagen1778 <roman@victoriametrics.com>
* lib/storage: follow-up after a50d63c376
- ensure retentionMsecs is rounded to day
- remove localTimeOffset in test as localOffset is ignored when using `UnixMilli`
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
* lib/storage: restore retention timezone offset effect on retention deadline
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
---------
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
previously during sync for mutable and immutable cache parts, link for hotEntry with current date may be not properly updated
it corrupts cache for backfilling metrics and increased cpu load
When using `retentionTimezoneOffset` and having local timezone being more than 4 hours different from UTC indexdb retention calculation could return negative value. This caused indexdb rotation to get in loop.
Fix calculation of offset to use `retentionTimezoneOffset` value properly and add test to cover all legit timezone configs.
See:
- https://github.com/VictoriaMetrics/VictoriaMetrics/issues/4207
- https://github.com/VictoriaMetrics/VictoriaMetrics/pull/4206
Signed-off-by: Zakhar Bessarab <z.bessarab@victoriametrics.com>
Co-authored-by: Nikolay <nik@victoriametrics.com>
Callers of OpenStorage() log the returned error and exit.
The error logging and exit can be performed inside MustOpenStorage()
alongside with printing the stack trace for better debuggability.
This simplifies the code at caller side.
Use fs.MustReadDir() instead of os.ReadDir() across the code in order to reduce the code verbosity.
The fs.MustReadDir() logs the error with the directory name and the call stack on error
before exit. This information should be enough for debugging the cause of the error.
Callers of CreateFlockFile log the returned err and exit.
It is better to log the error inside the MustCreateFlockFile together with the path
to the specified directory and the call stack. This simplifies
the code at the callers' side while leaving the debuggability at the same level.
Callers of InitFromFilePart log the error and exit.
It is better to log the error with the path to the part and the call stack
directly inside the MustInitFromFilePart() function.
This simplifies the code at callers' side while leaving the same level of debuggability.
Callers of this function log the returned error and exit.
It is better logging the error together with the path to the filename
and call stack directly inside the function. This simplifies
the code at callers' side without reducing the level of debuggability
Callers of this function log the returned error and exit.
Let's log the error with the path to the filename and call stack
inside the function. This simplifies the code at callers' side
without reducing the level of debuggability.
Callers of ReadFullData() log the error and then exit.
So let's log the error with the path to the filename and the call stack
inside MustReadData(). This simplifies the code at callers' side,
while leaving the debuggability at the same level.
Callers of these functions log the returned error and then exit.
Let's log the error with the call stack inside the function itself.
This simplifies the code at callers' side, while leaving the same
level of debuggability in case of errors.
Callers of this function log the returned error and then exit.
Let's log the error with the call stack inside the function itself.
This simplifies the code at callers' side, while leaving the same
level of debuggability in case of errors.
Callers of this function log the returned error and then exit.
Let's log the error with the call stack inside the function itself.
This simplifies the code at callers' side, while leaving the same
level of debuggability in case of errors.
Callers of this function log the returned error and exit.
So let's just log the error with the given filepath and the call stack
inside the function itself and then exit. This simplifies the code
at callers' place while leaves the same level of debuggability in case of errors.
Callers of these functions log the returned error and then exit. The returned error already contains the path
to directory, which was failed to be created. So let's just log the error together with the call stack
inside these functions. This leaves the debuggability of the returned error at the same level
while allows simplifying the code at callers' side.
While at it, properly use MustMkdirFailIfExist instead of MustMkdirIfNotExist inside inmemoryPart.MustStoreToDisk().
It is expected that the inmemoryPart.MustStoreToDick() must fail if there is already a directory under the given path.
When WriteFileAndSync fails, then the caller eventually logs the error message
and exits. The error message returned by WriteFileAndSync already contains the path
to the file, which couldn't be created. This information alongside the call stack
is enough for debugging the issue. So just use log.Panicf("FATAL: ...") inside MustWriteAndSync().
This simplifies error handling at caller side a bit.
This is a follow-up after 42bba64aa7
Previously the part directory listing was fsync'ed implicitly inside partHeader.WriteMetadata()
by calling fs.WriteFileAtomically(). Now it must be fsync'ed explicitly.
There is no need in fsync'ing the parent directory, since it is fsync'ed by the caller
when updating parts.json file.
Previously the created part directory listing was fsynced implicitly
when storing metadata.json file in it.
Also remove superflouous fsync for part directory listing,
which was called at blockStreamWriter.MustClose().
After that the metadata.json file is created, so an additional fsync
for the directory contents is needed.
Improperly configured -bigMergeConcurrency command-line flag usually leads to uncontrolled
growth of unmerged parts, which, in turn, increases CPU usage and query durations.
So it is better deprecating this flag. In rare cases -smallMergeConcurrency command-line flag
can be used instead for controlling the concurrency of background merges.
* lib/storage: check for free disk space before opening tables
We check for free disk space before call to `openTable`,
so `Storage` can be set to ReadOnly before mergeWorkers start.
Before the change, there was a chance that merges will start
even if Storage has to start in ReadOnly mode because of
`-storage.minFreeDiskSpaceBytes` limit.
https://github.com/VictoriaMetrics/VictoriaMetrics/issues/4023
Signed-off-by: hagen1778 <roman@victoriametrics.com>
* lib/storage: chore
Signed-off-by: hagen1778 <roman@victoriametrics.com>
* Update lib/storage/storage.go
---------
Signed-off-by: hagen1778 <roman@victoriametrics.com>
Co-authored-by: Aliaksandr Valialkin <valyala@victoriametrics.com>
- Use windows.FlushFileBuffers() instead of windows.Fsync() at streamTracker.adviseDontNeed()
for consistency with implementations for other architectures.
- Use filepath.Base() instead of filepath.Split(), since the dir part isn't used.
This simplifies the code a bit.
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/70
This is a follow-up for 43b24164ef
* lib/fs: adds memory map for windows
it should improve performance for file reading
* lib/storage: replace '/' with os specific separator
it must fix an errors for windows
* lib/fs: mention windows fsync support
* lib/filestream: adds fdatasync for windows writes
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/70
This commit changes background merge algorithm, so it becomes compatible with Windows file semantics.
The previous algorithm for background merge:
1. Merge source parts into a destination part inside tmp directory.
2. Create a file in txn directory with instructions on how to atomically
swap source parts with the destination part.
3. Perform instructions from the file.
4. Delete the file with instructions.
This algorithm guarantees that either source parts or destination part
is visible in the partition after unclean shutdown at any step above,
since the remaining files with instructions is replayed on the next restart,
after that the remaining contents of the tmp directory is deleted.
Unfortunately this algorithm doesn't work under Windows because
it disallows removing and moving files, which are in use.
So the new algorithm for background merge has been implemented:
1. Merge source parts into a destination part inside the partition directory itself.
E.g. now the partition directory may contain both complete and incomplete parts.
2. Atomically update the parts.json file with the new list of parts after the merge,
e.g. remove the source parts from the list and add the destination part to the list
before storing it to parts.json file.
3. Remove the source parts from disk when they are no longer used.
This algorithm guarantees that either source parts or destination part
is visible in the partition after unclean shutdown at any step above,
since incomplete partitions from step 1 or old source parts from step 3 are removed
on the next startup by inspecting parts.json file.
This algorithm should work under Windows, since it doesn't remove or move files in use.
This algorithm has also the following benefits:
- It should work better for NFS.
- It fits object storage semantics.
The new algorithm changes data storage format, so it is impossible to downgrade
to the previous versions of VictoriaMetrics after upgrading to this algorithm.
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3236
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/3821
Updates https://github.com/VictoriaMetrics/VictoriaMetrics/issues/70
