6e0c5698c1
Merge pull request #6766 from influxdata/nc-count-zero
...
Count and Sum return 0 for empty intervals
2016-06-02 08:45:08 -06:00
cd336095ca
Merge pull request #6768 from influxdata/jw-disable-open
...
Allow creating shards in a disabled state
2016-06-02 08:34:51 -06:00
ee4267bd05
Merge pull request #6767 from influxdata/jw-unlock
...
Fix sporadic write failures with influx_stress
2016-06-01 19:25:48 -06:00
579923d95f
Fix sporadic write failures with influx_stress
...
This Unlock was moved which seems to create a deadlock situation
sometimes under high write load. This deadlock causes writes to
fail with timeouts.
2016-06-01 17:25:47 -06:00
a74ea4cbf4
Allow creating shards in a disable state
...
For restoring a shard, we need to be able to have the shard open,
but disabled. It was racy to open it and then disable it separately
since writes/queries could occur in between that time.
2016-06-01 16:17:18 -06:00
ce74fe0b06
count and sum return 0 for empty intervals
2016-06-01 15:53:23 -06:00
71c8e9e567
Refactor ExecuteQuery to take options as a struct
...
This allows us to add additional options to ExecuteQuery without
creating parameter bloat.
Removing the unused Series structs. Their necessity was removed by a
previous commit, but the structs were not removed yet.
Add another type of interrupt iterator that monitors the interrupt
channel and calls `Close()` on the iterator when the interrupt happens.
It will primarily be used for asynchronously closing the ReaderIterator,
but it will only close the read side of the connection properly. More
work needs to be done to allow closing the write side efficiently.
2016-06-01 12:30:52 -05:00
1f390ee290
Merge pull request #6732 from influxdata/dn-batch-cq-writes2
...
batch SELECT INTO writes
2016-06-01 13:24:24 -04:00
ae9a57094e
Merge pull request #6763 from influxdata/er-monitor-dl
...
Fix deadlock in monitor
2016-06-01 18:09:54 +01:00
774e398612
Fix deadlock in monitor
2016-06-01 17:46:01 +01:00
62f3b4486a
batch SELECT INTO writes
2016-06-01 12:05:23 -04:00
05ec9599b7
Merge pull request #6744 from influxdata/jw-shard-enabled
...
Add ability to disable shards
2016-05-31 11:05:05 -06:00
d0023dee5d
Convert inline errors to constants
2016-05-31 10:51:54 -06:00
1ff8ecf4fb
Add ability to disable shards
...
Disabling a shard causes all writes and queries to a shard to return
an error. This also disables compactions for the shard.
2016-05-31 10:51:54 -06:00
baf5d505e6
Merge pull request #6754 from influxdata/er-fs
...
Prevent ReadFloatBlock from panicking when no values
2016-05-31 16:41:29 +01:00
003c30989a
Check for no values
2016-05-31 16:28:17 +01:00
2eaa12ef4b
Merge pull request #6745 from influxdata/fastpath-locking-in-cache-entries
...
Low-contention path for pre-existing cache entries.
2016-05-31 08:53:33 -06:00
dcec206f2e
Dedup `.RUnlock` between two conditionals.
2016-05-29 10:20:58 -07:00
1b160d1af0
Low-contention path for pre-existing cache entries.
...
This change appears to increase bulk ingestion throughput by 2x-3x in
multiprocessor environments.
2016-05-28 23:50:11 -07:00
dd58101061
Merge pull request #6743 from influxdata/jw-parse-key
...
Optimize series key parsing on startup
2016-05-27 15:00:42 -06:00
876c65acf8
Merge pull request #6741 from influxdata/jl-monitor-doc
...
Update comment for SetPointsWriter in the monitoring service
2016-05-27 15:31:18 -04:00
619c3c4229
Merge pull request #6735 from influxdata/jw-backup
...
Switch backup to use shard.Snapshot
2016-05-27 11:19:15 -06:00
ca13cb9932
Update changelog
2016-05-27 10:43:37 -06:00
ff1447202c
Reduce lock contention in Measurement.AddSeries
2016-05-27 10:30:08 -06:00
84c776bf2a
Skip allocating string in unescapeString
2016-05-27 10:30:08 -06:00
ff2475bf7c
Prevent allocation in unesecapeTag
2016-05-27 10:30:08 -06:00
97ad5fd2e6
Add ParseKey benchmark
2016-05-27 10:30:08 -06:00
f6ee765873
Merge pull request #6742 from influxdata/jw-utc-test
...
Use UTC for expected times in TestServer_Query_LargeTimestamp
2016-05-27 10:27:04 -06:00
0fb82d8787
Use UTC for expected times in TestServer_Query_LargeTimestamp
2016-05-27 09:46:07 -06:00
11959005f4
Switch backup to use shard.Snapshot
...
This switch the backup shard call to use the shard Snapshot that
internally creates a snapshot by hardlinking all of the TSM and
tombstone files instead. This reduces the time that the FileStore
is locked and will allow for larger shards to be backup more easily.
2016-05-27 09:30:25 -06:00
8c3ef8d8ae
Merge pull request #6730 from influxdata/jl-date-fix
...
Delay parsing of date/time strings until needed
2016-05-27 10:41:52 -04:00
381059a55c
Merge pull request #6736 from influxdata/benchmark-write-points-allocs
...
Benchmarks to count allocs in WritePoints.
2016-05-27 10:13:17 -04:00
f2fd988ab9
Delay parsing of date/time strings until needed
...
The current code would compare every string literal it crossed and tried
to coerce them to time literals if the _looked_ like date/time strings.
The only time the TimeLiteral was used is when comparing to the the
'time' value in a where clause. This change moves the string parsing
code until we attempt to compare 'time' to a string, at which point we
know we need/want a TimeLiteral, and not just an ordinary string.
Fixes #6727
2016-05-27 09:43:45 -04:00
da6a5ec7f8
Merge pull request #6722 from arussellsaw/master
...
mutex lock around measurement field map access
2016-05-27 14:06:42 +01:00
7edb14bffd
assign engine to shard after engine is initialized
2016-05-27 13:45:16 +01:00
f7c7b89b65
Merge pull request #6675 from influxdata/er-future-meta
...
Ensure meta SHOW queries include future points
2016-05-27 13:02:23 +01:00
b86c37ebeb
Merge pull request #6739 from influxdata/er-revert
...
Vet fix and revert
2016-05-27 12:43:20 +01:00
47942bc842
Update comment for SetPointsWriter in the monitoring service
2016-05-27 07:28:12 -04:00
2927fee2d1
update comment on MaxTime
2016-05-27 11:07:50 +01:00
9314ae8e80
fix overflow in window iterator and holt winters roundTime
2016-05-27 11:07:50 +01:00
f4fc905fa9
Reject timestamps too far in future
2016-05-27 11:07:48 +01:00
6a7f9527e3
Revert d2672a3 and 1e0a4e9
2016-05-27 10:34:14 +01:00
46489a5195
Fix vet issue
2016-05-27 10:31:20 +01:00
92e7fec5cf
Benchmarks to count allocs in WritePoints.
2016-05-26 17:13:14 -07:00
bc6e1e84af
Change max time for SHOW meta queries
...
If all points in a series are timestamped in the future, the SHOW
queries will not return anything from these series.
This commit changes the to time used when querying shards for the SHOW
queries to the maximum time, in order to ensure future points are
considered in the results for these queries.
Fixes #6599 .
2016-05-26 23:22:23 +01:00
6f25d97de4
Merge pull request #6677 from influxdata/er-ts-error
...
Ensure points with trailing whitespace are accepted
2016-05-26 22:36:13 +01:00
9f9aa88a4a
Merge pull request #6731 from influxdata/jl-monitor
...
Update monitor to allow setting the PointsWriter
2016-05-26 15:18:53 -04:00
39f3480f28
Ensure points with trailing whitespace are accepted
2016-05-26 19:00:24 +01:00
bd532f600f
Merge pull request #6692 from influxdata/ross-config-update
...
Remove cluster section from sample configuration
2016-05-26 13:57:41 -04:00
0317228487
Update monitor to allow setting the PointsWriter
...
The current code is very specific to clustering, which is no longer in
the open-source project. This changes the monitor service to allow for a
custom PointsWriter, rather than the remote-writing specific endpoint
that was there previously.
2016-05-26 12:36:44 -04:00
Nathaniel Cook
Jason Wilder
Jason Wilder
Nathaniel Cook
Jonathan A. Sternberg
David Norton
Edd Robinson
rw
joelegasse
Alex Russell-Saw