> "org.apache.solr.core.TestSolrDeletionPolicy1.testCommitAge fails almost > always on Windows" > > It hasn't been failing for me. I just ran the tests again. > > This is on Windows 7 Home Premium on an i5-based notebook PC. Maybe it is > sensitive to the particular environment or to some side effect of > virtualization.
Hi, I reproduced it (and other tests Solr also failing) with Windows on my Local machine running Windows 7 Prof 64bit, too. I think those errors may come from the fact that Windows does not allow writing/deleting files that are already open. I almost never run Solr tests on my windows machine, as it takes too long (horrible long), so I gave up on investigating this. With this Jenkins installation we are now able to handle that. I drilled down the windows tests to only run as a simple Jenkins Ant Build step; I will also make a copy of these Jobs on one of Apache's Windows slaves (if possible and works). > -- Jack Krupansky > > -----Original Message----- > From: Uwe Schindler > Sent: Tuesday, May 15, 2012 7:45 PM > To: dev@lucene.apache.org > Subject: New Jenkins for Windows and Linux > > Hi, > > I set up a new Jenkins server that runs Linux and Windows builds (both 64 > bit, running Java 7 and Java 6 latest): http://jenkins.sd-datasolutions.de > > The machine is running Ubuntu and has a 64 bit Windows 7 Professional > installation in a VirtualBOX VM managed by the Jenkins Master and some > local > scripts (Thanks to Garrett Serack for providing Apache with MSDN > licenses). > If new failure emails are coming, please look at them. Robert and me found > out: > > - The ongoing Solr test failures in TestReplicationHandler are also > happening (quite often on Linux, so definitely *not* an FreeBSD issue) > - org.apache.solr.core.TestSolrDeletionPolicy1.testCommitAge fails almost > always on Windows, I think the custom Deletion Policy of Solr does not > work > correctly on Windows > > You can see this as a test installation on my own company's (I am the > owner) > server until we have a dedicated Ubuntu slave running at Apache or in any > other data center. The setup of the windows-slave in a VirtualBox on Linux > was doable, but complicated, so I can help with that (the problem was not > VirtualBox, it was Windows 7 with its horrible security setting regarding > DCOM/Remote Registry access and User Group rights). The windows box has > now > 2 NICs, one using NAT to internet for IVY & Co and one private one only > for > communication between Jenkins master and slave, which is completely > private > (it passes all horrible DCOM/SAMBA stuff between master and slave). > Firewall > was turned off in Windows, otherwise it does not work (same applies to > spaces in windows user name -> bug in Jenkins :( ) > > Uwe > ----- > Uwe Schindler > H.-H.-Meier-Allee 63, D-28213 Bremen > http://www.thetaphi.de > eMail: u...@thetaphi.de > > > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org > For additional commands, e-mail: dev-h...@lucene.apache.org > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org > For additional commands, e-mail: dev-h...@lucene.apache.org > > --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org
