[
https://issues.apache.org/jira/browse/SOLR-6554?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14230004#comment-14230004
]
Shalin Shekhar Mangar commented on SOLR-6554:
---------------------------------------------
Here is one more test with 10 collections all having stateFormat=2
{code}
stateFormat = 2, 10 collections, With refactoring (trunk):
===========================================================
244536 T12 oasc.OverseerTest.testPerformance Overseer loop finished processing:
244541 T12 oasc.OverseerTest.printTimingStats totalTime: 230151.969647
244541 T12 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
0.004344859792294027
244542 T12 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute: 0.0
244542 T12 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute: 0.0
244542 T12 oasc.OverseerTest.printTimingStats avgTimePerRequest:
230151.969647
244543 T12 oasc.OverseerTest.printTimingStats medianRequestTime:
230151.969647
244543 T12 oasc.OverseerTest.printTimingStats 75thPctlRequestTime:
230151.969647
244543 T12 oasc.OverseerTest.printTimingStats 95thPctlRequestTime:
230151.969647
244543 T12 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
230151.969647
244544 T12 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
230151.969647
244544 T12 oasc.OverseerTest.testPerformance op: am_i_leader, success: 2,
failure: 0
244544 T12 oasc.OverseerTest.printTimingStats totalTime: 9.364505
244544 T12 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
0.5225137226108014
244545 T12 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute:
11.527327226155185
244545 T12 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute:
18.795347877757326
244545 T12 oasc.OverseerTest.printTimingStats avgTimePerRequest: 4.6822525
244545 T12 oasc.OverseerTest.printTimingStats medianRequestTime: 4.6822525
244545 T12 oasc.OverseerTest.printTimingStats 75thPctlRequestTime: 8.98267
244546 T12 oasc.OverseerTest.printTimingStats 95thPctlRequestTime: 8.98267
244546 T12 oasc.OverseerTest.printTimingStats 99thPctlRequestTime: 8.98267
244546 T12 oasc.OverseerTest.printTimingStats 999thPctlRequestTime: 8.98267
244546 T12 oasc.OverseerTest.testPerformance op: update_state, success: 20011,
failure: 0
244548 T12 oasc.OverseerTest.printTimingStats totalTime: 12520.257356
244549 T12 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
5230.413197388874
244549 T12 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute:
4052.517943905357
244550 T12 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute:
3016.875734960303
244550 T12 oasc.OverseerTest.printTimingStats avgTimePerRequest:
0.6256687499875069
244550 T12 oasc.OverseerTest.printTimingStats medianRequestTime:
0.5419605000000001
244551 T12 oasc.OverseerTest.printTimingStats 75thPctlRequestTime: 0.6782735
244551 T12 oasc.OverseerTest.printTimingStats 95thPctlRequestTime: 1.07258755
244551 T12 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
1.7485102400000028
244551 T12 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
8.31249523400001
244552 T12 oasc.OverseerTest.testPerformance op: state, success: 20001,
failure: 0
244554 T12 oasc.OverseerTest.printTimingStats totalTime: 14652.769701
244554 T12 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
5237.58427279227
244554 T12 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute:
4130.052867937011
244555 T12 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute:
3125.3532353495934
244555 T12 oasc.OverseerTest.printTimingStats avgTimePerRequest:
0.7326018549572522
244555 T12 oasc.OverseerTest.printTimingStats medianRequestTime: 0.6325495
244555 T12 oasc.OverseerTest.printTimingStats 75thPctlRequestTime: 0.771651
244556 T12 oasc.OverseerTest.printTimingStats 95thPctlRequestTime:
1.2006243999999988
244556 T12 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
2.0507515900000057
244556 T12 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
8.561998272000013
stateFormat = 2, 10 collections, Without refactoring (branch_5x):
=================================================================
1329102 T11 oasc.OverseerTest.testPerformance Overseer loop finished
processing:
1329105 T11 oasc.OverseerTest.printTimingStats totalTime: 1312114.128
1329106 T11 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
7.621263475999054E-4
1329107 T11 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute: 0.0
1329107 T11 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute: 0.0
1329107 T11 oasc.OverseerTest.printTimingStats avgTimePerRequest: 1312114.128
1329107 T11 oasc.OverseerTest.printTimingStats medianRequestTime: 1312114.128
1329108 T11 oasc.OverseerTest.printTimingStats 75thPctlRequestTime:
1312114.128
1329108 T11 oasc.OverseerTest.printTimingStats 95thPctlRequestTime:
1312114.128
1329108 T11 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
1312114.128
1329108 T11 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
1312114.128
1329108 T11 oasc.OverseerTest.testPerformance op: state, success: 20001,
failure: 0
1329110 T11 oasc.OverseerTest.printTimingStats totalTime: 466104.173
1329110 T11 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
915.6800211767887
1329111 T11 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute:
1418.8474273248714
1329111 T11 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute:
1043.4826929249386
1329111 T11 oasc.OverseerTest.printTimingStats avgTimePerRequest:
23.304043447827606
1329111 T11 oasc.OverseerTest.printTimingStats medianRequestTime: 2.3715
1329111 T11 oasc.OverseerTest.printTimingStats 75thPctlRequestTime:
2.7197500000000003
1329111 T11 oasc.OverseerTest.printTimingStats 95thPctlRequestTime:
3.977099999999994
1329112 T11 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
7.936150000000027
1329112 T11 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
35.16617800000001
1329112 T11 oasc.OverseerTest.testPerformance op: update_state, success: 20011,
failure: 0
1329114 T11 oasc.OverseerTest.printTimingStats totalTime: 327388.98
1329114 T11 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
915.6585782350854
1329114 T11 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute:
1425.935654658605
1329115 T11 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute:
1049.1858608633495
1329115 T11 oasc.OverseerTest.printTimingStats avgTimePerRequest:
16.360450752086354
1329115 T11 oasc.OverseerTest.printTimingStats medianRequestTime: 0.725
1329116 T11 oasc.OverseerTest.printTimingStats 75thPctlRequestTime: 0.844
1329116 T11 oasc.OverseerTest.printTimingStats 95thPctlRequestTime: 1.218
1329117 T11 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
2.1403600000000043
1329117 T11 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
12.222141000000011
1329117 T11 oasc.OverseerTest.testPerformance op: am_i_leader, success: 20013,
failure: 0
1329119 T11 oasc.OverseerTest.printTimingStats totalTime: 3622.9
1329120 T11 oasc.OverseerTest.printTimingStats avgRequestsPerMinute:
915.5821178966359
1329120 T11 oasc.OverseerTest.printTimingStats 5minRateRequestsPerMinute:
1424.927585830734
1329120 T11 oasc.OverseerTest.printTimingStats 15minRateRequestsPerMinute:
1042.4347274913655
1329121 T11 oasc.OverseerTest.printTimingStats avgTimePerRequest:
0.18102733223404788
1329121 T11 oasc.OverseerTest.printTimingStats medianRequestTime: 0.179
1329121 T11 oasc.OverseerTest.printTimingStats 75thPctlRequestTime: 0.197
1329122 T11 oasc.OverseerTest.printTimingStats 95thPctlRequestTime:
0.25909999999999994
1329122 T11 oasc.OverseerTest.printTimingStats 99thPctlRequestTime:
0.5761600000000034
1329123 T11 oasc.OverseerTest.printTimingStats 999thPctlRequestTime:
17.419063000000026
{code}
I'm not going to run more tests simply because they take too much time :)
> Speed up overseer operations for collections with stateFormat > 1
> -----------------------------------------------------------------
>
> Key: SOLR-6554
> URL: https://issues.apache.org/jira/browse/SOLR-6554
> Project: Solr
> Issue Type: Improvement
> Components: SolrCloud
> Affects Versions: 5.0, Trunk
> Reporter: Shalin Shekhar Mangar
> Attachments: SOLR-6554-batching-refactor.patch,
> SOLR-6554-batching-refactor.patch, SOLR-6554-batching-refactor.patch,
> SOLR-6554-batching-refactor.patch, SOLR-6554.patch, SOLR-6554.patch,
> SOLR-6554.patch, SOLR-6554.patch, SOLR-6554.patch, SOLR-6554.patch,
> SOLR-6554.patch, SOLR-6554.patch
>
>
> Right now (after SOLR-5473 was committed), a node watches a collection only
> if stateFormat=1 or if that node hosts at least one core belonging to that
> collection.
> This means that a node which is the overseer operates on all collections but
> watches only a few. So any read goes directly to zookeeper which slows down
> overseer operations.
> Let's have the overseer node watch all collections always and never remove
> those watches (except when the collection itself is deleted).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]