There are a number of code paths in Compactor.write which
on error can lead to leaked file handles to temporary files.
This, in turn, prevents the removal of the temporary files until
InfluxDB is rebooted, releasing the file handles.
closes https://github.com/influxdata/influxdb/issues/25724
Many tsm generations over level 4 compaction
single tsm generation under level 4 compaction all in
same shard. Group size is over 2 GB for each generation.
This PR adds an additional statistic "fluxQueryRespBytes" to the output of /debug/vars, in turn making it available to Telegraf and other monitoring tools.
Closes https://github.com/influxdata/influxdb/issues/25671
This PR changes the algorithm for compaction to account for the following
cases that were not previously accounted for:
- Many generations with a groupsize over 2 GB
- Single generation with many files and a groupsize under 2 GB
Where groupsize is the total size of the TSM files in said shard directory.
closes https://github.com/influxdata/influxdb/issues/25666
* fix(influxd): update xxhash, avoid stringtoslicebyte in cache (#578)
* fix(influxd): update xxhash, avoid stringtoslicebyte in cache
This commit does 3 things:
* it updates xxhash from v1 to v2; v2 includes a assembly arm version of
Sum64
* it changes the cache storer to write with a string key instead of a
byte slice. The cache only reads the key which WriteMulti already has
as a string so we can avoid a host of allocations when converting back
and forth from immutable strings to mutable byte slices. This includes
updating the cache ring and ring partition to write with a string key
* it updates the xxhash for finding the cache ring partition to use
Sum64String which uses unsafe pointers to directly use a string as a
byte slice since it only reads the string. Note: this now uses an
assembly version because of the v2 xxhash update. Go 1.22 included new
compiler ability to recognize calls of Method([]byte(myString)) and not
make a copy but from looking at the call sites, I'm not sure the
compiler would recognize it as the conversion to a byte slice was
happening several calls earlier.
That's what this change set does. If we are uncomfortable with any of
these, we can do fewer of them (for example, not upgrade xxhash; and/or
not use the specialized Sum64String, etc).
For the performance issue in maz-rr, I see converting string keys to
byte slices taking between 3-5% of cpu usage on both the primary and
secondary. So while this pr doesn't address directly the increased cpu
usage on the secondary, it makes cpu usage less on both which still
feels like a win. I believe these changes are easier to review that
switching to a byte slice pool that is likely needed in other places as
the compiler provides nearly all of the correctness checks we need (we
are relying also on xxhash v2 being correct).
* helps #550
* chore: fix tests/lint
* chore: don't use assembly version; should inline
This 2 line change causes xxhash to use a purego Sum64 implementation
which allows the compiler to see that Sum64 only read the byte slice
input which them means is can skip the string to byte slice allocation
and since it can skip that, it should inline all the calls to
getPartitionStringKey and Sum64 avoiding 1 call to Sum64String which
isn't inlined.
* chore: update ci build file
the ci build doesn't use the make file!!!
* chore: revert "chore: update ci build file"
This reverts commit 94be66fde03e0bbe18004aab25c0e19051406de2.
* chore: revert "chore: don't use assembly version; should inline"
This reverts commit 67d8d06c02e17e91ba643a2991e30a49308a5283.
(cherry picked from commit 1d334c679ca025645ed93518b7832ae676499cd2)
* feat: need to update go sum
---------
Co-authored-by: Phil Bracikowski <13472206+philjb@users.noreply.github.com>
(cherry picked from commit 06ab224516)
Add FutureWriteLimit and PastWriteLimit to retention
policies. Points which are outside of
now() + FutureWriteLimit
or
now() - PastWriteLimit
will be rejected on write with a PartialWriteError.
closes https://github.com/influxdata/influxdb/issues/25424
* chore: move shardID parsing and shard filtering into walkShardsAndProcess
* chore: make it impossible to miss sending shardResponse or marking shard as complete
* chore: always count number of shards (preparation for 2.x related feature)
* chore: explicitly load series files and create indices serially
Explicitly load series files and create indices serially. Also
avoid passing them to work functions that don't need them.
* chore: rework loadShards for changes necessary to cancel loading process
* chore: comment improvements
* fix: fix race conditions in TestStore_StartupShardProgress and TestStore_BadShardLoading
* chore: avoid logging nil error
* chore: refactor shard loading and shard walking
Refactor loadShards and CreateShard to use a common shardLoader class that
makes thread-safety easier. Refactor walkShardsAndProcess into findShards.
* chore: improve comment
* chore: rename OpenShard to ReopenShard and implement with shardLoader
Rename Store.OpenShard to Store.ReopenShard and implement using a
shardLoader object. Changes to tests as necessary.
* chore: avoid resetting shard options and locking on Reopen
Avoid resetting shard options when reopening a shard.
Proper mutex locker in Shard.ReopenShard.
* chore: fix formatting issue
* chore: warn on mixed index types in Store.CreateShard
* chore: change from info to warn when invalid shard IDs found in path
* chore: use coarser locking in Store.ReopenShard
* chore: fix typo in comment
* chore: code simplification
* feat(tsdb): Adds functionality to clear bad shards list
This PR adds test and new method to clear out the bad shards list
the method will return the values of the shards that it cleared out
along with the errors. This is the first part in the feature
for adding a load-shards command to influxd-ctl.
Closesinfluxdata/feature-requests#591
* feat(tsdb): Adds shard opening progress checks to startup
This PR adds a check to see how many shards are remaining
vs how many shards are opened. This change displays the percent
completed too.
closesinfluxdata/feature-requests#476
devanbenz
davidby-influx
cpinflux
Geoffrey Wossum
Shiwen Cheng