ankitsinghal edited a comment on issue #661: HBASE-15519 Add per-user metrics with lossy counting URL: https://github.com/apache/hbase/pull/661#issuecomment-541995255 Thanks @joshelser and @busbey for the review. Here is the result on ycsb run. ``` ./bin/ycsb load hbase20 -P workloads/workloada -p columnfamily=family -p recordcount=100000 ./bin/ycsb run hbase20 -P workloads/workloada -s -threads 10 -p columnfamily=family ``` Performance looks quite similar on 1 node cluster With patch ``` 2019-10-14 18:07:32:614 50 sec: 1000000 operations; 17966 current ops/sec; [READ: Count=2764, Max=8487, Min=125, Avg=547.82, 90=723, 99=990, 99.9=8027, 99.99=8487] [CLEANUP: Count=20, Max=1013, Min=0, Avg=71.4, 90=41, 99=1013, 99.9=1013, 99.99=1013] [READ-FAILED: Count=1725, Max=8047, Min=118, Avg=404.42, 90=506, 99=680, 99.9=1988, 99.99=8047] [UPDATE: Count=4460, Max=8583, Min=173, Avg=531.98, 90=656, 99=915, 99.9=8147, 99.99=8583] [OVERALL], RunTime(ms), 50503 [OVERALL], Throughput(ops/sec), 19800.803912638854 [TOTAL_GCS_PS_Scavenge], Count, 81 [TOTAL_GC_TIME_PS_Scavenge], Time(ms), 126 [TOTAL_GC_TIME_%_PS_Scavenge], Time(%), 0.24949012929924957 [TOTAL_GCS_PS_MarkSweep], Count, 1 [TOTAL_GC_TIME_PS_MarkSweep], Time(ms), 24 [TOTAL_GC_TIME_%_PS_MarkSweep], Time(%), 0.04752192939033325 [TOTAL_GCs], Count, 82 [TOTAL_GC_TIME], Time(ms), 150 [TOTAL_GC_TIME_%], Time(%), 0.2970120586895828 [READ], Operations, 268606 [READ], AverageLatency(us), 518.1957923501336 [READ], MinLatency(us), 125 [READ], MaxLatency(us), 46879 [READ], 95thPercentileLatency(us), 765 [READ], 99thPercentileLatency(us), 1230 [READ], Return=OK, 268606 [READ], Return=NOT_FOUND, 231808 [CLEANUP], Operations, 20 [CLEANUP], AverageLatency(us), 71.4 [CLEANUP], MinLatency(us), 0 [CLEANUP], MaxLatency(us), 1013 [CLEANUP], 95thPercentileLatency(us), 287 [CLEANUP], 99thPercentileLatency(us), 1013 [READ-FAILED], Operations, 231808 [READ-FAILED], AverageLatency(us), 413.0006168898399 [READ-FAILED], MinLatency(us), 118 [READ-FAILED], MaxLatency(us), 74559 [READ-FAILED], 95thPercentileLatency(us), 631 [READ-FAILED], 99thPercentileLatency(us), 1325 [UPDATE], Operations, 499586 [UPDATE], AverageLatency(us), 517.4709039084362 [UPDATE], MinLatency(us), 173 [UPDATE], MaxLatency(us), 92543 [UPDATE], 95thPercentileLatency(us), 737 [UPDATE], 99thPercentileLatency(us), 1426 [UPDATE], Return=OK, 499586 ``` without patch ``` 2019-10-14 18:21:37:869 50 sec: 1000000 operations; 18478.19 current ops/sec; [READ: Count=3520, Max=8179, Min=121, Avg=557.83, 90=731, 99=1317, 99.9=7931, 99.99=8179] [CLEANUP: Count=20, Max=866, Min=0, Avg=69.3, 90=35, 99=866, 99.9=866, 99.99=866] [READ-FAILED: Count=1941, Max=7907, Min=106, Avg=413.28, 90=543, 99=935, 99.9=2257, 99.99=7907] [UPDATE: Count=5508, Max=8151, Min=157, Avg=525.15, 90=648, 99=1314, 99.9=2385, 99.99=8139] [OVERALL], RunTime(ms), 50599 [OVERALL], Throughput(ops/sec), 19763.236427597385 [TOTAL_GCS_PS_Scavenge], Count, 71 [TOTAL_GC_TIME_PS_Scavenge], Time(ms), 122 [TOTAL_GC_TIME_%_PS_Scavenge], Time(%), 0.24111148441668806 [TOTAL_GCS_PS_MarkSweep], Count, 1 [TOTAL_GC_TIME_PS_MarkSweep], Time(ms), 30 [TOTAL_GC_TIME_%_PS_MarkSweep], Time(%), 0.05928970928279215 [TOTAL_GCs], Count, 72 [TOTAL_GC_TIME], Time(ms), 152 [TOTAL_GC_TIME_%], Time(%), 0.3004011936994802 [READ], Operations, 268278 [READ], AverageLatency(us), 516.6577617247781 [READ], MinLatency(us), 121 [READ], MaxLatency(us), 33983 [READ], 95thPercentileLatency(us), 760 [READ], 99thPercentileLatency(us), 1178 [READ], Return=OK, 268278 [READ], Return=NOT_FOUND, 231398 [CLEANUP], Operations, 20 [CLEANUP], AverageLatency(us), 69.3 [CLEANUP], MinLatency(us), 0 [CLEANUP], MaxLatency(us), 866 [CLEANUP], 95thPercentileLatency(us), 341 [CLEANUP], 99thPercentileLatency(us), 866 [READ-FAILED], Operations, 231398 [READ-FAILED], AverageLatency(us), 412.6978712002697 [READ-FAILED], MinLatency(us), 106 [READ-FAILED], MaxLatency(us), 94847 [READ-FAILED], 95thPercentileLatency(us), 612 [READ-FAILED], 99thPercentileLatency(us), 1263 [UPDATE], Operations, 500324 [UPDATE], AverageLatency(us), 520.759313964551 [UPDATE], MinLatency(us), 157 [UPDATE], MaxLatency(us), 116287 [UPDATE], 95thPercentileLatency(us), 730 [UPDATE], 99thPercentileLatency(us), 1370 [UPDATE], Return=OK, 500324 ```
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
