d474a0ed9c
Handle UnsignedLiteral in the Reduce operations
2017-09-27 15:10:30 -05:00
f0fb0174b6
Implement uint64 logic into eval
...
Any operation between an int64 and uint64 results in the type becoming a
uint64. This is because the most common will usually be an unsigned
cursor being modified by an int literal. Even in cases where we were to
add an unsigned literal to an integer iterator, that would just result
in it being recast back and overflow.
2017-09-27 15:10:16 -05:00
963cbfd275
Update eval type for unsigned
2017-09-27 11:31:52 -05:00
7fed382dbf
Merge pull request #8872 from influxdata/jw-mmap
...
Fix long process stalls
2017-09-25 14:49:35 -06:00
4e6075e819
Merge pull request #8873 from influxdata/js-emitter-unsigned-type
...
Support the unsigned type in query.Emitter
2017-09-25 15:45:28 -05:00
1d833cd920
Support the unsigned type in query.Emitter
2017-09-25 15:08:18 -05:00
122a74c692
Use synchronous IO for wal and tsm writing
...
The fysncs due to large writes when writing to TSM files and the
WAL can eventually cause large pauses. Since we already buffer
writes, using synchronous IO reduces fsync latency by ensuring
the individiual writes hit disk. This spreads out the latecncy
across multiple writes better.
2017-09-25 12:44:57 -06:00
660c2f8ae2
Merge pull request #8847 from influxdata/er-8842-engine-races
...
Fix various engine races
2017-09-25 18:34:29 +01:00
311f1fa619
Merge pull request #8864 from influxdata/js-8848-subquery-math-on-greater-than-two-fields
...
Prevent deadlock when doing math on the result of a subquery
2017-09-25 11:58:02 -05:00
612ed90b53
Merge pull request #8870 from nwante/patch-1
...
simple typo fix
2017-09-25 11:49:10 -05:00
2def219f09
Refactor Shard to further protect Engine
2017-09-25 17:43:30 +01:00
4a67f92acc
Prevent store from directly accessing Shard's engine
2017-09-25 17:43:01 +01:00
8e9cabbb9c
Fix race in TagValues when reaching into engine
2017-09-25 17:43:01 +01:00
7739ff749a
Ensure engine protected by shard mutex
2017-09-25 17:42:30 +01:00
2c43654552
Add some tests for concurrent server access
2017-09-25 17:42:30 +01:00
5774b44a4c
Remove MADV_RANDOM
...
This was inadvertently added when merging the solaris and unix
mmap files. This causes large delays due to major page faults.
2017-09-25 10:25:06 -06:00
58b674d16e
Merge pull request #8857 from influxdata/er-bloom
...
Improve performance of TSI Bloom Filter
2017-09-25 16:50:09 +01:00
e995bba24e
simple typo fix
...
just a simple wording fix when I was reading the docs.
2017-09-25 08:16:57 -04:00
e442bd87ea
Master 1657 ( #8865 )
...
Bugfix, was overwriting non-static package with static package
2017-09-22 16:59:54 -04:00
dbbb1090da
Merge pull request #8863 from influxdata/er-appveyor
...
Adjust AppVeyor CI options
2017-09-22 21:41:18 +01:00
bcf2e8fca5
Prevent deadlock when doing math on the result of a subquery
...
The `fill(none)` attribute got set on subqueries, but that can cause an
issue with certain subqueries just like it caused a deadlock on outer
queries.
2017-09-22 14:45:53 -05:00
9f3dcca512
Adjust AppVeyor CI options
...
The -race flag makes the tests timeout on the Windows VMs so we're
removing it. Race enabled tests will still happen for linux builds.
Also increases the timeout on the tests.
2017-09-22 18:30:38 +01:00
ea104596f0
Implement TSI index versioning
...
This commit adds a basic TSI versioning scheme, by adding a Version field
to an index's MANIFEST file.
Existing TSI indexes will not have this field present in their MANIFEST
files, and thus will be deemed incomatible with the current version.
Users with existing TSI indexes will be able to remove them, and convert the
resulting inmem indexes to the current version of a TSI index using the
influx_inspect tooling.
2017-09-22 17:59:39 +01:00
1028818ba6
Perf boost with some bit twiddling
2017-09-22 17:59:39 +01:00
5b7fc517fa
Improve performance of TSI bloom filter
...
This commit replaces the previous hashing algorithm used by the pkg.Filter with
one based on xxhash. Further, taking from the hashing literature, we can
represent k hashes with only two hash function, where previously Filter was using
four.
Further, unlike `murmur3`, `xxhash` is allocation-free, so allocations have
dramatically reduced when inserting and checking for hashes.
2017-09-22 17:59:39 +01:00
fe960b0f3a
Add benchmarks for bloom filter
2017-09-22 17:59:32 +01:00
1e345aa7a1
Merge pull request #8856 from influxdata/jw-cache
...
Snapshot compaction improvements
2017-09-22 10:45:54 -06:00
ccb6f7451f
Ensure importer emits errors to stderr ( #8833 )
2017-09-22 17:17:20 +01:00
e0caa2e0fd
Merge pull request #8696 from influxdata/er-8678-tsi1-where
...
Allow tag filtering when using DELETE with tsi1
2017-09-22 17:16:09 +01:00
44691847e9
Merge branch 'master' into er-8678-tsi1-where
2017-09-22 16:54:49 +01:00
80aa328be6
small bugfix in filename check for 'static' keywork ( #8861 )
2017-09-22 10:33:12 -04:00
94aba64b88
Re-use index entries slice when writing TSM index
2017-09-21 12:48:16 -06:00
db204f3eb7
Default concurrent compactions to 50% of available cores
2017-09-21 12:48:11 -06:00
5eb79f438e
Merge pull request #8852 from influxdata/rb-8816-glyphicons
...
Remove unused glyphicons package from licenses file.
2017-09-21 12:19:36 -04:00
fba7c9fc5e
Updated packaging script to better handle static packaging ( #8859 )
2017-09-20 14:04:01 -04:00
deef0c5649
Fix 32bit alignment
2017-09-20 10:00:20 -06:00
61ca1243c7
Increase index disk writer buffer
2017-09-20 09:05:30 -06:00
796de3dcea
Reduce encoder pool checkout contention
...
With higher cardinalities, the encoder pools where become a bottleneck.
This changes the snapshot compactions ot checkout one encoder of each
type and re-use it while writing the snapshots as opposed to repeatedly
checking it out and in.
2017-09-19 15:27:26 -06:00
391a6288c6
Write parallel snapshot for higher cardinalities
2017-09-19 15:27:26 -06:00
0d52b060df
Skip onFileStoreReplace with tsi
2017-09-19 15:27:25 -06:00
4fe81aeee6
Remove manual Gosched from compactions
...
At higher cardinalities, this dramatically slows down compaction throughput.
2017-09-19 15:27:25 -06:00
31e785d676
Don't deduplicate a single value
2017-09-19 15:27:25 -06:00
2ca9ccee1f
Reset snapshot cache outside of write lock
2017-09-19 15:27:25 -06:00
ddeba2c86b
Split large snapshots and write concurrently
2017-09-19 15:27:25 -06:00
9ee305f6f5
Periodically re-allocate cache store
...
This perioically re-allocates the cache store to avoid memory
fragmentation and gradual slow down of the store after repeated
deletes and inserts into the map.
2017-09-19 15:27:25 -06:00
2885b9b310
Remove entrySizeHints map
...
There is a lot of overhead for calculating the hints for larger
cardinalities. This slows down resetting the partitions in the ring.
2017-09-19 15:27:25 -06:00
4124a8ed97
Simplify cache ring
...
The continuum slice is not needed since the number of partitions
doesn't change. This removes the slice to make the mapping simpler.
2017-09-19 15:27:25 -06:00
65bcddf6c0
Merge pull request #8855 from influxdata/sgc-fix-findrange
...
fix FindValues panic for empty array
2017-09-19 14:26:45 -07:00
ed7bc9d825
fix FindValues panic for empty array
2017-09-19 14:23:32 -07:00
cf0a1d9d99
Merge pull request #8853 from influxdata/sgc-readentries-perf
...
Reduce allocations, improve readEntries performance by simplifying loop
2017-09-19 13:28:43 -07:00
Jonathan A. Sternberg
Jason Wilder
Jason Wilder
Edd Robinson
nwante
Adam
Ryan Betts
Stuart Carnie