loop devices again - fedora18-vm02 needs a reboot

[Fabian Deutsch]

Am Dienstag, den 13.08.2013, 17:19 +0200 schrieb David Caro:
> On Tue 13 Aug 2013 02:21:23 PM CEST, Kiril Nesenko wrote:
> > Done
> >
> > - Kiril
> >
> > ----- Original Message -----
> >> From: "Fabian Deutsch" <fabiand at redhat.com>
> >> To: infra at ovirt.org
> >> Sent: Tuesday, August 13, 2013 3:10:08 PM
> >> Subject: loop devices again -  fedora18-vm02  needs a reboot
> >>
> >> Hey,
> >>
> >> could someone restart  fedora18-vm02 there are many orphaned loop
> >> devices, see [1].
> >>
> >> Thanks
> >> fabian
> >>
> >>
> >> --
> >> [1]
> >> http://jenkins.ovirt.org/view/ovirt_node/job/node-devel/distro=fedora18/142/consoleFull
> >>
> >> _______________________________________________
> >> Infra mailing list
> >> Infra at ovirt.org
> >> http://lists.ovirt.org/mailman/listinfo/infra
> >>
> > _______________________________________________
> > Infra mailing list
> > Infra at ovirt.org
> > http://lists.ovirt.org/mailman/listinfo/infra
> 
> That machine has now 256 loop devices available, but I think that it's 
> not a solution. The job that got the loop devices busy was node-devel 
> (http://jenkins.ovirt.org/view/vdsm/job/node-devel/).
> 
> I see it has a somehow complicated config scirpt but I think that the 
> problem is in the make scripts, I see this in the logs on all the jobs:
> 
> Loop device does not match a floppy size, using default hd params
> ERROR:program:losetup: /dev/loop7: detach failed: Device or resource 
> busy
> losetup: /dev/loop7: detach failed: Device or resource busy
> Initialized /dev/loop8 as a 19 MB HFS Plus volume
> ERROR:program:losetup: /dev/loop8: detach failed: Device or resource 
> busy
> losetup: /dev/loop8: detach failed: Device or resource busy
> ERROR:program:losetup: /dev/loop9: detach failed: Device or resource 
> busy
> losetup: /dev/loop9: detach failed: Device or resource busy
> Traceback (most recent call last):
>   File "/sbin/mkefiboot", line 153, in <module>
>     opt.diskname)
>   File "/sbin/mkefiboot", line 45, in mkmacboot
>     macmunge(outfile, product)
>   File "/sbin/mkefiboot", line 60, in macmunge
>     shim = glob.glob(os.path.join(mnt, 'EFI/BOOT/BOOT*.efi'))[0]
> IndexError: list index out of range
> 
> Anyone is familiar with this job? (I'm not sure if the IndexError is 
> related to the loop devices error though)

Hey,

yes - I guess it was me who caused the problems, because I stopped some
ovirt-node jobs through the jenkins interface. I didn't know that the
ovirt-node job now also contains the iso part (the job got reconfigured
while I was on vacation) - because stopping that, killed livecd creator
which further lead to the orphaned loop devices.

I think it's now at least the third time that we discuss the problems
with ovirt-node - or more specififc with the livecd-creator.
Besides ovirt-node, it's also the job which creates the ovirt-live iso
which can lead to similar problems.
We won't be able to sort out all problems we see as long as we use
livecd-creator. We hope that it will get better when we migrate to
imagefac (which will be using VMs to create image).

Greetings
fabian

> 
> --
> David Caro
> 
> Red Hat Czech s.r.o.
> Continuous Integration Engineer - EMEA ENG Virtualization R&D
> 
> Tel.: +420 532 294 605
> Email: dcaro at redhat.com
> Web: www.cz.redhat.com
> Red Hat Czech s.r.o., Purkyňova 99/71, 612 45, Brno, Czech Republic
> RHT Global #: 82-62605
> _______________________________________________
> Infra mailing list
> Infra at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/infra