Becoming an sub-project infra member

[Fabian Deutsch]

Am Freitag, den 19.04.2013, 04:53 -0400 schrieb Eyal Edri:
> 
> ----- Original Message -----
> > From: "Fabian Deutsch" <fabiand at redhat.com>
> > To: "Eyal Edri" <eedri at redhat.com>
> > Cc: infra at ovirt.org
> > Sent: Friday, April 19, 2013 11:47:09 AM
> > Subject: Re: Becoming an sub-project infra member
> > 
> > Hey Eyal,
> > 
> > Am Freitag, den 19.04.2013, 04:33 -0400 schrieb Eyal Edri:
> > > 
> > > ----- Original Message -----
> > > > From: "Fabian Deutsch" <fabiand at redhat.com>
> > > > To: infra at ovirt.org
> > > > Sent: Thursday, April 18, 2013 7:16:05 PM
> > > > Subject: Re: Becoming an sub-project infra member
> > > > 
> > > > Am Dienstag, den 16.04.2013, 15:33 +0200 schrieb Fabian Deutsch:
> > > > > Hey,
> > > > > 
> > > > > I'm Fabian and I'm working on the oVirt Node sub-project. The special
> > > > > nature of Node leads to a tight coupling between our build system and
> > > > > the host. Because of that I'd like to become a (co-)maintainer of build
> > > > > hosts were oVirt Node is build. To look after the host side packages
> > > > > and
> > > > > to be in a better position to debug build problems.
> > > > > 
> > > > > Is something ike this possible?
> > > > 
> > > > Anyone if I can get access to the F18 slaves?
> > > > 
> > > 
> > > why do you need access to the slaves? and what kind of access
> > > (root/sudo/read only?)
> > > please specify the requirements.
> > 
> > The main reason is to keep the livecd-tools/selinux packages up to date
> > and to initiate a slave restart once all loop devices are exhausted.
> 
> this can be done via a 'maintanance jenkins job' i assume, if you can write a script
> that does it, we can give it the right sudo access to run it periodically or manually.

That's a good idea! Basically only a reboot is the reliable way to
remove orphaned loop devices.
It would be along the lines:

echo "Checking for orphan loop devices"
[[ $(/usr/sbin/losetup -a | grep "(deleted)" | wc -l) -lt 0 ]] && {
  echo "Orphaned loop devices detected, initiating reboot"
  rpc_call_to_disable_renkins_slave (?)
  reboot
}

But could a jenkins slave temporarily disabled like this (from within a
job)?

> > Additionally we are seeing build errors which we can not reproduce on
> > our local builders - which I'd also like to debug.
> 
> i think this requires read only access - ewoud/quaid - what is our policy for giving ssh access to jenkins slaves for RO?
> 
> > 
> > What I'd need would be the permission to temporarily disable slaves for
> > rebooting and local access to reboot a slave.
> > 
> > > there is a job in jenkins that installs pkg on the slaves, i can give you
> > > permissions to it
> > > if needed w/o need to access the slaves.
> > 
> > That will at least address the problem to have up to date packages.
> > 
> > Greetings
> > fabian
> > 
> > 
> > > eyal.
> > > 
> > > > Thanks
> > > > fabian
> > > > 
> > > > > Thanks
> > > > > fabian
> > > > > 
> > > > > 
> > > > > _______________________________________________
> > > > > 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
> > > > 
> > 
> > 
> >