Re: random failure on tests

2013-05-29 Thread Nicolas Liochon
Yes for the stack size. See as well the config 2.1.2.5, nofile & nproc are important for the tests. Check as well with the resource checker (see chapter 15.7.3.7) that the values actually used in the tests execution are the ones that you expect. You want MaxFileDescriptor to be above 2K to be safe.

Re: random failure on tests

2013-05-29 Thread Lisen Mu
Nicolas, Thanks for the reply! I run small & medium tests only, and I have 8G ram on the build server. My passing rate is also around 50%, but I'm not running -PrunAllTests. I can find 2 differences between official build and mine: missing prebuild step 'rm -rf /tmp/hbase-jenkins/hbase' and stac

Re: random failure on tests

2013-05-29 Thread Nicolas Liochon
Hello, Option 1: We still have some flaky tests. You can benchmark you build against https://builds.apache.org/job/HBase-TRUNK/ and https://builds.apache.org/job/hbase-0.95/ You can also use this tool: https://github.com/jeffreyz88/jenkins-tools to get a review on the last fails: On 0.95, we have

random failure on tests

2013-05-28 Thread Lisen Mu
Hello, I'm setting up a jenkins job for hbase, building branch 0.95 (from github) under jdk 6. However sometimes the build pasts, sometimes does not. Several tests are likely to fail, such as: org.apache.hadoop.hbase.ipc.TestDelayedRpc.testDelayedRpcImmediateReturnValue org.apache.hadoop.hbase.i