Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

tune grpc connection count between tidb and tikv #12884

Merged
merged 2 commits into from
Oct 23, 2019

Conversation

zhangjinpeng87
Copy link
Contributor

Signed-off-by: zhangjinpeng1987 zhangjinpeng@pingcap.com

What problem does this PR solve?

Tune the gRPC connection between TiDB and TiKV to make super batch works better. This is a prepare step of tikv/tikv#5598, this pr will make tikv/tikv#5598 works better.

What is changed and how it works?

Tune the default gRPC connection between tidb and tikv.

Check List

Tests

  • Unit test

Signed-off-by: zhangjinpeng1987 <zhangjinpeng@pingcap.com>
@zhangjinpeng87
Copy link
Contributor Author

/bench

@sre-bot
Copy link
Contributor

sre-bot commented Oct 22, 2019

@@                               Benchmark Diff                               @@
================================================================================
--- tidb: 076503a4683f91e40ec8f1c14119ed532e6fda63
+++ tidb: 738da0ed2148fad3b34309f0aeebd78f7d30583e
tikv: 453e866835d654e20ae4d01fe75ba17abf8577f9
pd: 752ecd784fabfe1841a252be29c01284de8098f6
================================================================================
oltp_update_non_index:
    * QPS: 4844.92 ± 0.46% (std=13.94) delta: -0.39% (p=0.187)
    * Latency p50: 26.42 ± 0.46% (std=0.07) delta: 0.38%
    * Latency p99: 44.60 ± 4.56% (std=1.45) delta: 4.66%
            
oltp_insert:
    * QPS: 4687.65 ± 0.18% (std=6.10) delta: -0.04% (p=0.853)
    * Latency p50: 27.30 ± 0.18% (std=0.04) delta: 0.04%
    * Latency p99: 50.34 ± 2.24% (std=0.75) delta: -0.74%
            
oltp_read_write:
    * QPS: 15115.50 ± 0.05% (std=5.61) delta: 0.23% (p=0.878)
    * Latency p50: 169.73 ± 0.06% (std=0.08) delta: -0.19%
    * Latency p99: 320.17 ± 0.00% (std=0.00) delta: -1.78%
            
oltp_update_index:
    * QPS: 4334.35 ± 0.53% (std=13.56) delta: 0.36% (p=0.144)
    * Latency p50: 29.47 ± 0.07% (std=0.01) delta: -0.44%
    * Latency p99: 56.08 ± 2.23% (std=0.83) delta: -3.14%
            

@codecov
Copy link

codecov bot commented Oct 22, 2019

Codecov Report

Merging #12884 into master will increase coverage by 0.0074%.
The diff coverage is n/a.

@@               Coverage Diff                @@
##             master     #12884        +/-   ##
================================================
+ Coverage   79.9822%   79.9897%   +0.0074%     
================================================
  Files           465        465                
  Lines        107255     106940       -315     
================================================
- Hits          85785      85541       -244     
+ Misses        15028      14960        -68     
+ Partials       6442       6439         -3

@tabokie
Copy link

tabokie commented Oct 22, 2019

/bench

tidb.yml:
  tikv_client:
    # Max gRPC connections that will be established with each tikv-server.
    grpc-connection-count: 4

@sre-bot
Copy link
Contributor

sre-bot commented Oct 22, 2019

@@                               Benchmark Diff                               @@
================================================================================
--- tidb: 5ea98344c482f325847a5c1e2ba0daf1759c1600
+++ tidb: 738da0ed2148fad3b34309f0aeebd78f7d30583e
tikv: 453e866835d654e20ae4d01fe75ba17abf8577f9
pd: 1af456e3b493dc1e9de5ef85ab7595909b4d9796
================================================================================
oltp_update_non_index:
    * QPS: 4866.21 ± 0.41% (std=12.26) delta: 0.31% (p=0.225)
    * Latency p50: 26.30 ± 0.41% (std=0.07) delta: -0.31%
    * Latency p99: 42.05 ± 3.19% (std=0.99) delta: -3.13%
            
oltp_insert:
    * QPS: 4686.11 ± 0.39% (std=12.92) delta: 0.20% (p=0.383)
    * Latency p50: 27.31 ± 0.37% (std=0.07) delta: -0.20%
    * Latency p99: 50.11 ± 0.00% (std=0.00) delta: -1.35%
            
oltp_point_select:
    * QPS: 35341.56 ± 0.89% (std=216.30) delta: 6.83% (p=0.105)
    * Latency p50: 3.62 ± 0.90% (std=0.02) delta: -6.31%
    * Latency p99: 12.75 ± 1.80% (std=0.16) delta: -1.39%
            

@zhangjinpeng87 zhangjinpeng87 requested review from disksing and breezewish and removed request for disksing October 23, 2019 01:28
@tabokie tabokie self-requested a review October 23, 2019 01:46
@zhangjinpeng87 zhangjinpeng87 merged commit 7447361 into pingcap:master Oct 23, 2019
@zhangjinpeng87 zhangjinpeng87 deleted the grpc-connection branch October 23, 2019 01:47
lfkdsk added a commit to JustProject/tidb that referenced this pull request Oct 26, 2019
…ect/tidb into feature-add-udf-support

* 'feature-add-udf-support' of https://github.com/JustProject/tidb: (26 commits)
  *: fix bug that the kill command doesn't work when the killed session is waiting for the pessimistic lock (pingcap#12852)
  executor: fix the projection upon the indexLookUp in indexLookUpJoin can't get result. (pingcap#12889)
  planner, executor: support create view on union (pingcap#12595)
  planner/cascades: introduce TransformationID in cascades planner (pingcap#12879)
  executor: fix data race in test (pingcap#12910)
  executor: reuse chunk row for insert on duplicate update (pingcap#12847)
  ddl: speed up tests (pingcap#12888)
  executor: speed up test (pingcap#12896)
  expression: implement vectorized evaluation for `builtinSecondSig` (pingcap#12886)
  expression: implement vectorized evaluation for `builtinJSONObjectSig` (pingcap#12663)
  expression: speed up builtinRepeatSig by using MergeNulls (pingcap#12674)
  expression: speed up unit tests under the expression package (pingcap#12887)
  store,kv: snapshot doesn't cache the non-exists kv entries lead to poor 'insert ignore' performance (pingcap#12872)
  executor: fix data race in `GetDirtyTable()` (pingcap#12767)
  domain: increase TTL to reduce the occurrence of reporting min startTS errors (pingcap#12578)
  executor: split test for speed up (pingcap#12881)
  executor: fix inconsistent of grants privileges with MySQL when executing `grant all on ...` (pingcap#12330)
  expression: implement vectorized evaluation for `builtinJSONUnquoteSig` (pingcap#12841)
  tune grpc connection count between tidb and tikv (pingcap#12884)
  Makefile: change test parallel to 8 (pingcap#12885)
  ...
XiaTianliang pushed a commit to XiaTianliang/tidb that referenced this pull request Dec 21, 2019
Signed-off-by: zhangjinpeng1987 <zhangjinpeng@pingcap.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants