
Adding Evgheni. On Wed, Apr 19, 2017 at 10:01 AM, Nadav Goldin <ngoldin@redhat.com> wrote:
Hi Milan, sorry for missing this.
In short, it looks like a libvirt/qemu error, I guess it lays somewhere in the nested environment the Jenkins slave runs at. I was able to extract the libvirt log from this specific run, but there is nothing useful there, except that there was no proper termination. From reading here[1] it might be related to a load on the hypervisor, and the timeout configured for libvirt to wait for qemu. Unfortunately looking at the this[2] thread, it seems that a patch to configure the timeout never got into libvirt, which leaves us with a default of 30 seconds, and that might not be enough in our nested environment. I presume that if the hypervisor which the Jenkins slave runs is highly loaded, then when we try to start the vdsm_functional_tests_lago VM, it might take more than 30 seconds for qemu to respond.
Another indication of this "hypothesis" is that I never seen this error on OST - which uses bare-metal slaves.
Evgheni, do we have the load monitoring on the hypervisor that runs vm0065.workers-phx.ovirt.org? Not sure if we added that eventually.
[1] https://bugzilla.redhat.com/show_bug.cgi?id=987088 [2] https://www.redhat.com/archives/libvir-list/2014-January/msg00410.html
Hi,
after my Vdsm patch https://gerrit.ovirt.org/75329 in ovirt-4.1 branch had been merged, Jenkins check-merged job http://jenkins.ovirt.org/job/vdsm_4.1_check-merged-el7-x86_64/173/ failed with the following error:
07:01:21 @ Start specified VMs: 07:01:21 # Start nets: 07:01:21 * Create network vdsm_functional_tests_lago: 07:01:27 * Create network vdsm_functional_tests_lago: Success (in 0:00:05) 07:01:27 # Start nets: Success (in 0:00:05) 07:01:27 # Start vms: 07:01:27 * Starting VM vdsm_functional_tests_host-el7: 07:02:07 libvirt: QEMU Driver error : monitor socket did not show up: No such file or directory 07:02:07 * Starting VM vdsm_functional_tests_host-el7: ERROR (in 0:00:40) 07:02:07 # Start vms: ERROR (in 0:00:40) 07:02:07 # Destroy network vdsm_functional_tests_lago: 07:02:07 # Destroy network vdsm_functional_tests_lago: ERROR (in 0:00:00) 07:02:07 @ Start specified VMs: ERROR (in 0:00:46) 07:02:07 Error occured, aborting 07:02:07 Traceback (most recent call last): 07:02:07 File "/usr/lib/python2.7/site-packages/lago/cmd.py", line 936, in main 07:02:07 cli_plugins[args.verb].do_run(args) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/plugins/cli.py",
07:02:07 self._do_run(**vars(args)) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/utils.py",
07:02:07 return func(*args, **kwargs) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/utils.py",
07:02:07 return func(*args, prefix=prefix, **kwargs) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/cmd.py", line 264, in do_start 07:02:07 prefix.start(vm_names=vm_names) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/prefix.py",
07:02:07 self.virt_env.start(vm_names=vm_names) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/virt.py", line 331, in start 07:02:07 vm.start() 07:02:07 File "/usr/lib/python2.7/site-packages/lago/plugins/vm.py",
07:02:07 return self.provider.start(*args, **kwargs) 07:02:07 File "/usr/lib/python2.7/site-packages/lago/vm.py", line 106, in start 07:02:07 dom = self.libvirt_con.createXML(self._libvirt_xml()) 07:02:07 File "/usr/lib64/python2.7/site-packages/libvirt.py", line 3782, in createXML 07:02:07 if ret is None:raise libvirtError('virDomainCreateXML() failed', conn=self) 07:02:07 libvirtError: monitor socket did not show up: No such file or
On Mon, Apr 10, 2017 at 10:56 AM, Milan Zamazal <mzamazal@redhat.com> wrote: line 184, in do_run line 495, in wrapper line 506, in wrapper line 1033, in start line 299, in start directory
07:02:07 Took 210 seconds
The error is apparently unrelated to my patch since: 1. my patch should have nothing to do with VM start; 2. Jenkins has run successfully on the following patch (https://gerrit.ovirt.org/75321). FWIW, the preceding patch (https://gerrit.ovirt.org/75038) has run successfully too.
Do you know what's wrong?
Thanks, Milan _______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
-- Eyal edri ASSOCIATE MANAGER RHV DevOps EMEA VIRTUALIZATION R&D Red Hat EMEA <https://www.redhat.com/> <https://red.ht/sig> TRIED. TESTED. TRUSTED. <https://redhat.com/trusted> phone: +972-9-7692018 irc: eedri (on #tlv #rhev-dev #rhev-integ)