It also looks like TestMasterFailover isn't cleaning up after itself.
Each test method just does TEST_UTIL.shutdownMiniCluster() at the end,
without use of a finally block. So it wouldn't get called in the case
of an exception. Testing a patch for that now.
On Wed, Nov 23, 2011 at 9:52 AM, Gar
>
> Shall we upgrade hadoop version for hadoop-0.20 to 0.20.205.1-7070-SNAPSHOT
> as well ?
>
This is an unreleased version that's only necessary for ZooKeeper SASL
authentication using jaas.conf (needs HADOOP-7070).
I think it's better for the core build to remain on a released version
of Hadoop
I think this might be caused by:
https://issues.apache.org/jira/browse/HBASE-4857
There is a recursive loop on the leader
AuthenticationTokenSecretManger when a KeeperException is encountered.
On Wed, Nov 23, 2011 at 8:53 AM, Ted Yu wrote:
> The heap space error happened again for build #9
>
>
https://issues.apache.org/jira/browse/HBASE-4856 has been created.
Shall we upgrade hadoop version for hadoop-0.20 to 0.20.205.1-7070-SNAPSHOT
as well ?
On Wed, Nov 23, 2011 at 9:00 AM, Stack wrote:
> On Wed, Nov 23, 2011 at 8:53 AM, Ted Yu wrote:
> > The heap space error happened again for bu
On Wed, Nov 23, 2011 at 8:53 AM, Ted Yu wrote:
> The heap space error happened again for build #9
>
> TestMasterFailover hung in build 9 too.
>
Yes. You mean the pom setting? Want to make an issue Ted w/ patch?
(Maybe more crap going on when security in the mix pushing is over
edge; we should d
The heap space error happened again for build #9
TestMasterFailover hung in build 9 too.
On Tue, Nov 22, 2011 at 2:07 PM, Ted Yu wrote:
> The interesting thing is that their counterparts in HBase-0.92 build
> didn't run out of heap :-)
>
>
> On Tue, Nov 22, 2011 at 2:05 PM, Todd Lipcon wrote:
The interesting thing is that their counterparts in HBase-0.92 build didn't
run out of heap :-)
On Tue, Nov 22, 2011 at 2:05 PM, Todd Lipcon wrote:
> Usually IME this means that some test is generating many GB of logs,
> etc... before raising heap let's look and see if some of the tests are
> be
Usually IME this means that some test is generating many GB of logs,
etc... before raising heap let's look and see if some of the tests are
being excessively noisy
-Todd
On Tue, Nov 22, 2011 at 2:01 PM, Ted Yu wrote:
> Hi,
> I found the following in
> https://builds.apache.org/view/G-L/view/HBas
Hi,
I found the following in
https://builds.apache.org/view/G-L/view/HBase/job/HBase-0.92-security/lastCompletedBuild/console
:
Running org.apache.hadoop.hbase.master.TestDistributedLogSplitting
Exception in thread "ThreadedStreamConsumer"
java.lang.OutOfMemoryError: Java heap space
at jav