Test failures on lucene_solr_3_3 and branch_3x

[Shawn Heisey <so...@elyograg.org>]

I've got a consistent test failure on Solr source code checked out from 
svn.  The same thing happens with 3.3 and branch_3x.  I have information 
saved from the failures on branch_3x, which I have gotten to to fail 
about a dozen times in a row.  It fails on a test called 
TestSqlEntityProcessorDelta, part of the dataimporthandler tests.  It is 
consistently reproducible in a shorter timeframe than normal with the 
following commandline:

ant test -Dtestcase=TestSqlEntityProcessorDelta

Comprehensive ant output here, from a full test run:

http://pastebin.com/eyAt8Qg8

Platform information:

[root@idxst0-a solr]# uname -a
Linux idxst0-a 2.6.18-238.12.1.el5.centos.plusxen #1 SMP Wed Jun 1 
11:57:54 EDT 2011 x86_64 x86_64 x86_64 GNU/Linux
[root@idxst0-a solr]# cat /etc/redhat-release
CentOS release 5.6 (Final)
[root@idxst0-a solr]# java -version
java version "1.6.0_26"
Java(TM) SE Runtime Environment (build 1.6.0_26-b03)
Java HotSpot(TM) 64-Bit Server VM (build 20.1-b02, mixed mode)

[root@idxst0-a yum.repos.d]# yum repolist
Loaded plugins: fastestmirror, protectbase
Loading mirror speeds from cached hostfile
  * addons: mirror.san.fastserv.com
  * base: mirrors.tummy.com
  * centosplus: mirror.san.fastserv.com
  * contrib: mirror.san.fastserv.com
  * epel: mirrors.xmission.com
  * extras: mirrors.xmission.com
  * jpackage-generic: jpackage.netmindz.net
  * jpackage-generic-nonfree: www.mirrorservice.org
  * jpackage-generic-nonfree-updates: www.mirrorservice.org
  * jpackage-generic-updates: jpackage.netmindz.net
  * jpackage-rhel: jpackage.netmindz.net
  * jpackage-rhel-updates: jpackage.netmindz.net
  * rpmforge: fr2.rpmfind.net
  * updates: mirrors.tummy.com

Re: Test failures on lucene_solr_3_3 and branch_3x

[Shawn Heisey <so...@elyograg.org>]

On 9/2/2011 10:25 AM, Shawn Heisey wrote:
>> if you do: can you tell us more about the filesystem you are using?
> I am building this on an NFSv3 filesystem.  The NFS server is a 
> Solaris 10 cluster, the underlying filesystem is ZFS, on a 
> fiberchannel SAN.  I will check out the tree onto a local ext3 
> filesystem and try it again.

Still a consistent failure, two different systems, not failing on the 
new assumeTrue lines you added.  When it's not on NFS, it sure builds a 
lot faster. :)

The first one is the same as before, it's CentOS 5.6 running in a Xen 
VM, ext3 filesystem.  The second one is CentOS 6, 
2.6.18-238.12.1.el5.centos.plusxen, no virtualization, ext4 on LVM.  
Oracle java version is 1.6.0_26 64bit on both.  On the CentOS 5.6 
machine, ant and junit come from jpackage 5, on CentOS 6 I have base, 
contrib, centosplus, epel, extras, rpmforge, and rpmforge-extras.

Thanks,
Shawn

Re: Test failures on lucene_solr_3_3 and branch_3x

[Shawn Heisey <so...@elyograg.org>]

On 9/1/2011 4:12 PM, Chris Hostetter wrote:
> : I've got a consistent test failure on Solr source code checked out from svn.
> : The same thing happens with 3.3 and branch_3x.  I have information saved from
>
> Shawn: sorry for hte late reply.
>
> I can't reproduce your specific problem, but the test in question is
> suspiciously hinky enough taht i took a stab at making it work a little
> cleaner...
>
> 	https://issues.apache.org/jira/browse/SOLR-2739
>
> ..can you try running "ant test -Dtestcase=TestSqlEntityProcessorDelta"
> a few times on the updated 3x branch and comment in the issue about the
> results.
>
> particularly: do you get any failures? if not, do you get any messages in
> the log about asumptions failing?
>
> if you do: can you tell us more about the filesystem you are using?

I replied to the Jira issue yesterday, but didn't see this until today.

The failure looks to be the same, failing on this line:

       assertQ(req("id:1"), "//*[@numFound='0']");

I am building this on an NFSv3 filesystem.  The NFS server is a Solaris 
10 cluster, the underlying filesystem is ZFS, on a fiberchannel SAN.  I 
will check out the tree onto a local ext3 filesystem and try it again.

Thanks,
Shawn

Re: Test failures on lucene_solr_3_3 and branch_3x

[Chris Hostetter <ho...@fucit.org>]

: I've got a consistent test failure on Solr source code checked out from svn.
: The same thing happens with 3.3 and branch_3x.  I have information saved from

Shawn: sorry for hte late reply.

I can't reproduce your specific problem, but the test in question is 
suspiciously hinky enough taht i took a stab at making it work a little 
cleaner...

	https://issues.apache.org/jira/browse/SOLR-2739

..can you try running "ant test -Dtestcase=TestSqlEntityProcessorDelta" 
a few times on the updated 3x branch and comment in the issue about the 
results.

particularly: do you get any failures? if not, do you get any messages in 
the log about asumptions failing?

if you do: can you tell us more about the filesystem you are using?

-Hoss