This updates influxql to a newer version that supports complex bound
parameters. This allows bound parameters to be used as identifiers,
regexes, and durations along with the already existing strings, numbers,
and booleans.
This updates the influxdb client to support passing bound parameters as
part of the parameters json and updates the readme to show how to use
this feature.
We were seing segfaults in Roaring bitmaps sometimes, under very
high load with networked drives. This may reduce risk of segfault by
forcing marshalling to copy the data.
This commit adds a /api/v2/write endpoint that maps the supplied bucket
and org to a v1 database and retention policy.
* Add AllowedOrgs to httpd Config type.
* Add /api/v2/write handler
* fix: access tsi active log file with READ lock
The activeLogFile pointer may be altered by other routine so the READ
lock is needed.
* Merge pull request #16384 from foobar/tsi-partition-lock
fix: access tsi active log file with READ lock
Co-authored-by: Tristan Su <sooqing@gmail.com>
Co-authored-by: David Norton <dgnorton@gmail.com>
* Add .mod and .sum files generated by go mod init
* Use a directory outside of $GOPATH for testing.
* Update docker files to use go 1.13.8 instead of 1.12 for most builds.
* Simplify go installation by piping from wget to tar. This avoids
having to do any cleanup.
* Build and test in a directory outside of GOPATH since we're using
modules now.
* Update Jenkinsfile to build using modules by building directly inside
of $WORKSPACE instead of in $GOPATH.
* Use go mod download instead of dep ensure to collect dependencies
* Remove Gopkg related data files.
* Replace calls to "dep ensure" with "go mod download"
* Replace check for being outside of $GOPATH to a check for being INSIDE
of $GOPATH which is wrong when we're using modules
* Map $DIR to /root/influxdb in docker run command
* Clone source into /root/influxdb which is outside of our $GOPATH.
* run go mod tidy
When an InfluxDB database is very busy writing new points the backup
the process can fail because it can not write a new snapshot.
The error is: `operation timed out with error: create snapshot: snapshot in progress`.
This happens because InfluxDB takes almost "continuously" a snapshot
from the cache caused by the high number of points ingested.
This PR skips snapshots if the `snapshotter` does not come available
after three attempts when a backup is requested.
The backup won't contain the data in the cache or WAL.
Signed-off-by: Gianluca Arbezzano <gianarb92@gmail.com>
refactor(influxdb): Refactor trace code for clarity and reliability
* Make startProfile a method of Server. startProfile() is only used one
place. We bother to copy the Server.CPUProfile and Server.MemProfile
values out of our Server struct into it's parameters. It makes more
sense to make startProfile() a method of Server and have it access
those members directory via its receiver value.
* Have startProfile() return an error instead of log.Fatal()ing. We can
simply propagate the error up the stack and let the caller handle the
error -- we shouldn't be exiting deep in the bowels of a non-main
package.
* Capture and return errors from pprof.StartCPUProfile(). Currently
there is only one possible error it can return but if it returns an
error, we should handle it.
* add CPUProfileWriteCloser and MemProfileWriteCloser to Server struct.
* make stopProfile() a method of Server
* remove prof variable
* fix captialization of log messages.
This PR lazily initializes Flux built-in functions when Flux is used.
It significantly reduces the startup time of the `influxd` and `influx`
binaries.
Before (4.66s):
```
↳ time bin/18/influx
bin/18/influx 4.66s user 0.19s system 198% cpu 2.441 total
```
After (10ms):
```
↳ time bin/18/influx
bin/18/influx 0.01s user 0.01s system 88% cpu 0.021 total
```
Prior to this change, new series would be added to the series file
before checking the series cardinality limit. If the limit was exceeded,
the write was rejected even though the series had already been added to
the series file.
This commit prevents multiple blocks for the same series key having
values truncated when they are being read into an empty buffer.
The current cursor reader code has an optimisation that incorrectly
assumes the incoming array will be limited to 1,000 values (the maximum
block size), but arrays can contain values from multiple matching
blocks.
* fix(storage): skip TSM files with block read errors
When we find a bad TSM file during compaction, propagate the error up and move
the bad file aside. The engine will disregard the file so the next compaction
will not hit the same error.
Jonathan A. Sternberg
jsign
David Norton
docmerlin (j. Emrys Landivar)
Ayan George
Edd Robinson
Stuart Carnie
Ben Johnson
Sebastian Borza
Hercules Mango Churchill
Gianluca Arbezzano
Gianluca Arbezzano
Brandon Farmer
Sean Brickley
ricky
tmgordeeva
Jacob Marble
Edd Robinson