Tracked in SOLR-5948 so we don't lose the logs due ot jenkins job expiration...
https://issues.apache.org/jira/browse/SOLR-5948 : Date: Wed, 2 Apr 2014 09:56:19 -0700 (MST) : From: Chris Hostetter <[email protected]> : To: [email protected] : Subject: Re: [JENKINS] Lucene-Solr-trunk-Windows (32bit/jdk1.7.0_51) - Build # : 3917 - Failure! : : : This is a new test (committed yesterday) but it does't create/write/modify : any files on disk directly -- just using the regular distributed test : setup. : : Can anyone with a windows machine reproduce this failure to cleanup the : workspace? : : : 1 tests failed. FAILED: : : junit.framework.TestSuite.org.apache.solr.cloud.DistribDocExpirationUpdateProcessorTest : : : : Error Message: !!!! WARNING: best effort to remove : : C:\Users\JenkinsSlave\workspace\Lucene-Solr-trunk-Windows\solr\build\solr-core\test\J0\.\solrtest-DistribDocExpirationUpdateProcessorTest-1396449841389 : : FAILED !!!!! : : Possibly related is that at one point one node attempts to do log replay : and encounteres a "FileNotFoundException: _0.si" ... no idea why/how that : could happen, but it definitley smells fishy. And of course: right after : the FNFE start, we start seeing lots of 500 errors from other nodes trying : to communicate with this node whose index is missing files. : : : : : (FWIW: There's also an NPE getting logged here indicating that there is no : Leader at one point with the processor runs. That NPE shouldn't have : caused any sort of problem -- it's in an islated Runable that logs the : error and ignores -- but i'll fix the code to be more careful. In : anycase, the lack of a leader at that point may be indicitive of some : underlying problem that also then lead to the FNFE and ultimatley the : cleanup failure? : : : : -Hoss : http://www.lucidworks.com/ : -Hoss http://www.lucidworks.com/ --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
