[
https://issues.apache.org/jira/browse/ZOOKEEPER-1146?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13082887#comment-13082887
]
Hadoop QA commented on ZOOKEEPER-1146:
--------------------------------------
-1 overall. Here are the results of testing the latest attachment
http://issues.apache.org/jira/secure/attachment/12489411/ZOOKEEPER-1146.patch
against trunk revision 1152141.
+1 @author. The patch does not contain any @author tags.
-1 tests included. The patch doesn't appear to include any new or modified
tests.
Please justify why no new tests are needed for this
patch.
Also please list what manual steps were performed to
verify this patch.
+1 javadoc. The javadoc tool did not generate any warning messages.
+1 javac. The applied patch does not increase the total number of javac
compiler warnings.
+1 findbugs. The patch does not introduce any new Findbugs (version 1.3.9)
warnings.
+1 release audit. The applied patch does not increase the total number of
release audit warnings.
+1 core tests. The patch passed core unit tests.
+1 contrib tests. The patch passed contrib unit tests.
Test results:
https://builds.apache.org/job/PreCommit-ZOOKEEPER-Build/439//testReport/
Findbugs warnings:
https://builds.apache.org/job/PreCommit-ZOOKEEPER-Build/439//artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
Console output:
https://builds.apache.org/job/PreCommit-ZOOKEEPER-Build/439//console
This message is automatically generated.
> significant regression in client (c/python) performance
> -------------------------------------------------------
>
> Key: ZOOKEEPER-1146
> URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1146
> Project: ZooKeeper
> Issue Type: Bug
> Components: c client
> Affects Versions: 3.4.0
> Reporter: Patrick Hunt
> Assignee: Patrick Hunt
> Priority: Blocker
> Fix For: 3.4.0
>
> Attachments: ZOOKEEPER-1146.patch
>
>
> I tried running my latency tester against trunk, in so doing I noticed that
> the C/Python (not sure which yet) client performance has seriously degraded
> since 3.3.3.
> The first run (below) is with released 3.3.3 client against a 3 server
> ensemble running released 3.3.3 server code. The second run is the exact same
> environment (same ensemble), however using trunk c/zkpython client.
> Notice:
> 1) in the first run operations are approx 10ms/write, 0.25ms/read - which is
> pretty much what's expected.
> 2) however in the second run we are seeing 50ms/operation regardless of read
> or write.
> {noformat}
> [phunt@c0309 zk-smoketest-3.3.3]$ PYTHONPATH=lib.linux-x86_64-2.6/
> LD_LIBRARY_PATH=lib.linux-x86_64-2.6/ python26 ./zk-latencies.py --servers
> "c0309:2181,c0310:2181,c0311:2181" --znode_size=100 --znode_count=100
> --timeout=5000 --synchronous
> Connecting to c0309:2181
> Connected in 16 ms, handle is 0
> Connecting to c0310:2181
> Connected in 16 ms, handle is 1
> Connecting to c0311:2181
> Connected in 15 ms, handle is 2
> Testing latencies on server c0309:2181 using syncronous calls
> created 100 permanent znodes in 959 ms (9.599378 ms/op 104.173415/sec)
> set 100 znodes in 933 ms (9.332101 ms/op 107.157002/sec)
> get 100 znodes in 27 ms (0.270889 ms/op
> 3691.551589/sec)
> deleted 100 permanent znodes in 881 ms (8.812950 ms/op 113.469388/sec)
> created 100 ephemeral znodes in 956 ms (9.564152 ms/op 104.557103/sec)
> watched 100 znodes in 26 ms (0.264361 ms/op
> 3782.707587/sec)
> deleted 100 ephemeral znodes in 881 ms (8.819292 ms/op 113.387792/sec)
> notif 100 watches in 999 ms (9.994299 ms/op 100.057038/sec)
> Testing latencies on server c0310:2181 using syncronous calls
> created 100 permanent znodes in 964 ms (9.640460 ms/op 103.729490/sec)
> set 100 znodes in 933 ms (9.332800 ms/op 107.148981/sec)
> get 100 znodes in 29 ms (0.299308 ms/op
> 3341.036650/sec)
> deleted 100 permanent znodes in 886 ms (8.864651 ms/op 112.807603/sec)
> created 100 ephemeral znodes in 958 ms (9.585140 ms/op 104.328161/sec)
> watched 100 znodes in 30 ms (0.300801 ms/op
> 3324.459240/sec)
> deleted 100 ephemeral znodes in 886 ms (8.865030 ms/op 112.802779/sec)
> notif 100 watches in 1000 ms (10.000212 ms/op 99.997878/sec)
> Testing latencies on server c0311:2181 using syncronous calls
> created 100 permanent znodes in 958 ms (9.582071 ms/op 104.361569/sec)
> set 100 znodes in 935 ms (9.359350 ms/op 106.845024/sec)
> get 100 znodes in 25 ms (0.252700 ms/op
> 3957.263893/sec)
> deleted 100 permanent znodes in 891 ms (8.913291 ms/op 112.192013/sec)
> created 100 ephemeral znodes in 958 ms (9.584489 ms/op 104.335246/sec)
> watched 100 znodes in 25 ms (0.251091 ms/op
> 3982.627356/sec)
> deleted 100 ephemeral znodes in 891 ms (8.915379 ms/op 112.165730/sec)
> notif 100 watches in 1000 ms (10.000508 ms/op 99.994922/sec)
> Latency test complete
> [phunt@c0309 zk-smoketest-3.3.3]$ cd ../zk-smoketest-trunk/
> [phunt@c0309 zk-smoketest-trunk]$ PYTHONPATH=lib.linux-x86_64-2.6/
> LD_LIBRARY_PATH=lib.linux-x86_64-2.6/ python26 ./zk-latencies.py --servers
> "c0309:2181,c0310:2181,c0311:2181" --znode_size=100 --znode_count=100
> --timeout=5000 --synchronous
> Connecting to c0309:2181
> Connected in 31 ms, handle is 0
> Connecting to c0310:2181
> Connected in 16 ms, handle is 1
> Connecting to c0311:2181
> Connected in 16 ms, handle is 2
> Testing latencies on server c0309:2181 using syncronous calls
> created 100 permanent znodes in 5099 ms (50.999281 ms/op 19.608119/sec)
> set 100 znodes in 5066 ms (50.665429 ms/op 19.737324/sec)
> get 100 znodes in 4009 ms (40.093150 ms/op 24.941916/sec)
> deleted 100 permanent znodes in 5040 ms (50.404449 ms/op 19.839519/sec)
> created 100 ephemeral znodes in 5124 ms (51.249170 ms/op 19.512511/sec)
> watched 100 znodes in 4051 ms (40.514441 ms/op 24.682557/sec)
> deleted 100 ephemeral znodes in 5048 ms (50.484939 ms/op 19.807888/sec)
> notif 100 watches in 1000 ms (10.004182 ms/op 99.958199/sec)
> Testing latencies on server c0310:2181 using syncronous calls
> created 100 permanent znodes in 5115 ms (51.157510 ms/op 19.547472/sec)
> set 100 znodes in 5056 ms (50.568910 ms/op 19.774996/sec)
> get 100 znodes in 4099 ms (40.999382 ms/op 24.390612/sec)
> deleted 100 permanent znodes in 5041 ms (50.418010 ms/op 19.834182/sec)
> created 100 ephemeral znodes in 5083 ms (50.835850 ms/op 19.671157/sec)
> watched 100 znodes in 4100 ms (41.003261 ms/op 24.388304/sec)
> deleted 100 ephemeral znodes in 5058 ms (50.581930 ms/op 19.769906/sec)
> notif 100 watches in 1000 ms (10.005081 ms/op 99.949219/sec)
> Testing latencies on server c0311:2181 using syncronous calls
> created 100 permanent znodes in 5099 ms (50.992720 ms/op 19.610642/sec)
> set 100 znodes in 5091 ms (50.916569 ms/op 19.639972/sec)
> get 100 znodes in 4099 ms (40.996401 ms/op 24.392385/sec)
> deleted 100 permanent znodes in 5066 ms (50.669601 ms/op 19.735699/sec)
> created 100 ephemeral znodes in 5124 ms (51.249208 ms/op 19.512496/sec)
> watched 100 znodes in 4099 ms (40.999141 ms/op 24.390755/sec)
> deleted 100 ephemeral znodes in 5049 ms (50.498819 ms/op 19.802443/sec)
> notif 100 watches in 999 ms (9.997852 ms/op 100.021486/sec)
> Latency test complete
> {noformat}
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
