* 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)
* chore: upgrade Go to 1.19.3
This re-runs ./generate.sh and ./checkfmt.sh to format and update
source code (this is primarily responsible for the huge diff.)
* fix: update tests to reflect sorting algorithm change
* feat: upgrade flux to 0.171.0
Tests failing, safety commit
First step in https://github.com/influxdata/influxdb/issues/23815
* fix: remove "org" parameter" from writeOptSource
I attempted to implement the "orgOpt" argument in a similar fashion
to f6669f7512. However, it looks like Flux doesn't accept "org" as
a parameter to "load". It responds with:
Error calling function \"load\" @113:16-113:30: error calling function \"to\" @6:19-6:47: unused arguments [org]
This brings us from 194 passing to 570 passing.
* fix: temporarily disable broken flux tests
These tests expect rows to be stored in a certain order. However,
nothing is specifying the sort order. This has been fixed in a
later update to flux: (see 3d6f47ded).
Temporarily disable these tests until we include a fixed
version of the flux tests.
* chore: add tests from a492993012
This fixes "test-flux.sh" so it runs tests within the "flux/"
directory. This uncovered some other issues with the tests
located within "flux/". These also needed to be updated
to match the newer flux API.
* feat: upgrade flux to 0.172.0
This includes changes made in "cbbf4b27da". Since "test.go" in 2.x
diverged from 1.x, some modifications were required to make this
compatible.
* feat: upgrade flux to 0.173.0
* feat: upgrade flux to v0.174.0
* fix: Update the condition when reseting cursor (#23522)
Filters that contain `or` may change between cursor resets so we must remember to update the condition in the read cursor.
```flux
|> filter(fn: (r) => ((r["_field"] == "field1" and r["_value"]==true) or (r["_field"] == "field2" and r["_value"] == false)))
```
Closes https://github.com/influxdata/flux/issues/4804
* feat: upgrade flux to 0.174.1
* feat: upgrade flux to 0.175.0
* chore: remove end-to-end tests
These were removed in a492993 for 2.x. These tests prevent "go test ./..."
from completing. As stated in the original commit, these tests should now be
handled by the "fluxtest" harness.
* feat: upgrade flux to 0.176.0
Some tests needed to be disabled within the flux harness. This is a
result of enabling "Optimize Aggregate Window" in flux@05a1065f.
These tests are not present in 2.x. Therefore, I am unsure if
the breakage is resolved in a later commit.
* feat: upgrade flux to 0.177.0
* feat: upgrade flux to 0.178.0
* feat: upgrade flux to v0.179.0
This removes all invocations of "flux.RegisterOpSpec". According
to flux@e39096d5, "flux.RegisterOpSpec" does nothing in the
current version of flux and was removed.
* chore: update fluxtest skip list (#23633)
* chore: manually backport 785a465e9a
This removes the reference to "flux.Spec".
* build(flux): update flux to v0.181.0 (#23682)
* build(flux): update flux to v0.184.2
* chore: skip more Flux acceptance tests
There are issues for each skip detailed in test-flux.sh.
* feat: upgrade flux to v0.185.0
This adds "FluxTesting" to the "HTTPD" configuration. This option is
hidden and disabled by default. When "FluxTesting" is set, it
enables the default testing flags for "Flux".
These flags allow the "vectorized float tests" and tests requiring
the "removeRedundantSortNodes" and "labelPolymorphism" flag
enabled to work. These changes are based off of d8553c002e.
flux@3d6f47ded is included within this version of Flux. Therefore
we can now include the "group_*" tests.
* feat: upgrade flux to 0.186.0
* feat: upgrade flux to 0.187.0
* feat: upgrade flux to 0.188.0
* fix: re-run ./generate.sh with updated protoc
* fix: restrict cores to match CircleCI documentation
Co-authored-by: davidby-influx <dbyrne@influxdata.com>
Co-authored-by: Markus Westerlind <marwes91@gmail.com>
Co-authored-by: Sean Brickley <sean@wabr.io>
Co-authored-by: Jonathan A. Sternberg <jonathan@influxdata.com>
Co-authored-by: Christopher M. Wolff <chris.wolff@influxdata.com>
When a file rename fails with EXDEV
(cross device or volume error), copy the
file and delete the original instead
Differs from master branch by overwriting
existing files instead of erring.
closes https://github.com/influxdata/influxdb/issues/22997
* chore: pull in unchanged kit from v2
* chore: remove v2 from import paths
* chore: update module paths and go.mod for kit
* chore: remove kit/cli again, not needed in 1.x
* feat(query): hyper log log counting in query engine
In addition to helping with normal queries, this can improve the 'SHOW CARDINALITY'
meta-queries:
time influx -database mydb -execute 'select count_hll(sum_hll(_seriesKey)) from big'
name: big
time count_hll
---- ---------
0 200767781
influx -database mydb -execute 0.06s user 0.12s system 0% cpu 8:49.99 total
When comparing strings in a case-insensitive way, strings.EqualFold() is
(almost?) always faster than comparing the results of strings.ToLower().
In addition, strings.EqualFold() never causes an allocation.
This patch replaces case-insensitive string comparisons that use
strings.ToLower() with a strings.EqualFold() call.
This commit quiets staticcheck's warnings about "unnecessary use of
fmt.Sprintf" and "unnecessary use of fmt.Sprint".
Prior to this commit we were wrapping simple constant strings without
any formatting verbs with fmt.Sprintf().
This integrates the influxdb 1.x series to the latest version of Flux
and updates the code to use it. It also removes the dependency on
platform and copies the necessary code from storage into the 1.x series
so the dependency is unneeded.
The flux functions specific to 1.x have been moved to the same structure
that flux changed to with having a `stdlib` directory instead of a
`functions` directory. It also adds a `databases()` function that
returns the databases from the meta client.
Specifically:
* renamed files for consistency between versions
* added `time-interval` schema option
* updated schema example documentation
Back port of improvements from #12710
If 120th or 240th value is not a 1, k still passes the check in the
switch, causing the last value to be lost. If this value occurs at
the boundary of a block, the max time will be incorrect, resulting in
compaction failing to make forward progress.
it's slightly slower, but the safety is worth it i think.
```
name old time/op new time/op delta
Next-8 30.0ns ± 2% 31.0ns ± 3% +3.56% (p=0.002 n=7+8)
NextParallel-8 79.4ns ± 1% 92.5ns ± 1% +16.58% (p=0.000 n=8+8)
```
use atomics rather than mutexes to synchronize state between calls.
```
name old time/op new time/op delta
Next-8 244ns ± 0% 30ns ± 2% -87.70% (p=0.000 n=8+7)
NextParallel-8 215µs ±60% 0µs ± 1% -99.96% (p=0.000 n=8+8)
```
The results for NextParallel are around ~80ns/op, but that doesn't
show up in the benchstat output.
WeblWabl
davidby-influx
Brandon Pfeifer
Jakub Bednář
Dane Strandboge
Dane Strandboge
Sam Arnold
Ayan George
Jacob Marble
Jonathan A. Sternberg
Jeff Wendling
Stuart Carnie
Jeff Wendling
Ben Johnson
KimMachineGun
Ben Wells
liukun
Jeff Wendling