Agreed on both counts - I'd already updated the message to say that as well as supporting the standard "recheck" you can run "citrix recheck" to only trigger a recheck of XenServer and will try and fix the http headers tomorrow.
Bob ________________________________________ From: David Kranz [dkr...@redhat.com] Sent: 23 February 2014 21:17 To: Bob Ball; #OpenStack External Email Cc: OpenStack Development Mailing List Subject: Re: [qa] Can't interpret failure from XenServer CI On 02/23/2014 03:29 PM, Bob Ball wrote: > Hi David, > > That's a very strange error - it basically means the node has been used for > running a test already (both the first error in run_tests.log about the git > repository also existing, and the more fatal error about vif 3 existing) > > I believe this was my fault - I should have purged all negative results > before allowing the CI to post them, or at least re-tested them. > > You're right in the comment that "citrix recheck" will only re-trigger this > test and it is currently being tested. It is currently running and therefore > a proper result should be available soon. > > I will re-queue all of the "Failed" jobs so that a new XS testing result will > be posted overnight. > > Bob Thanks Bob. In addition to my comments about the log arrangement and providing http headers to make downloading .txt.gz work, it would be helpful to have the review comment mention 'citrix recheck'. I only discovered that through a web search. In general, the more we can do to establish uniformity of reviewer experience across all third-party ci, the more likely it is that reviewers will examine failures. -David > ________________________________________ > From: David Kranz [dkr...@redhat.com] > Sent: 23 February 2014 20:06 > To: #OpenStack External Email > Cc: OpenStack Development Mailing List > Subject: [qa] Can't interpret failure from XenServer CI > > XenServer reported failure on https://review.openstack.org/#/c/73704. > The pointer to the logs should look like a jenkins failure but does not > in two ways. First, there is no console log, with a directory of other > log files next to it. Second, the .gz log files are not set up to be > downloaded as text as they are in the jenkins link. I took a guess that > http://ca.downloads.xensource.com/OpenStack/xenserver-ci/refs/changes/04/73704/14/run_tests.log > might contain the console output. It has a bunch of errors that > obviously have nothing to do with the changed code. > > What are reviewers supposed to do when seeing such a failure? > > -David _______________________________________________ OpenStack-dev mailing list OpenStack-dev@lists.openstack.org http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
