[ovirt-devel] [VDSM] Build failure because missing ioprocess package in ovirt repos - solved

[Francesco Romani]

----- Original Message -----
> From: "Sandro Bonazzola" <sbonazzo at redhat.com>
> To: "Nir Soffer" <nsoffer at redhat.com>
> Cc: "Piotr Kliczewski" <pkliczew at redhat.com>, "devel" <devel at ovirt.org>
> Sent: Friday, August 19, 2016 10:55:28 AM
> Subject: Re: [ovirt-devel] [VDSM] Build failure because missing ioprocess package in ovirt repos - solved
> 
> 
> 
> On Thu, Aug 18, 2016 at 6:03 PM, Sandro Bonazzola < sbonazzo at redhat.com >
> wrote:
> 
> 
> 
> 
> 
> On Thu, Aug 18, 2016 at 4:16 PM, Nir Soffer < nsoffer at redhat.com > wrote:
> 
> 
> Merged, so I sent this to enable the ppc builds again:
> https://gerrit.ovirt.org/62551
> 
> Someone with access to jenkins should verify that this jobs works now.
> 
> 
> 
> will check tomorrow morning
> 
> 
> Looks ok to me

Hi all,

after a chat with Sandro (thanks to him for pointing this out),
it looks to me that we have a similar issue for Vdsm.

We could use the same fix as ioprocess (https://gerrit.ovirt.org/#/c/62464/), it seems
simple and safe - or at least I can't think of any problem atm.

Would everyone be OK with this?

(please scroll down [quote intentionally kept] for more context)

> On Thu, Aug 18, 2016 at 10:03 AM, Sandro Bonazzola < sbonazzo at redhat.com >
> wrote:
> > 
> > 
> > On Wed, Aug 17, 2016 at 6:21 PM, Nir Soffer < nsoffer at redhat.com > wrote:
> >> 
> >> I posted this patch, changing python-ioprocess to require
> >> ioprocess-version (instead of version-release). This seem to resolve
> >> the issue and is more correct, we don't really need ioprocess from the
> >> same build.
> >> 
> >> Please review
> >> https://gerrit.ovirt.org/62464
> > 
> > 
> > +1
> > 
> > 
> >> 
> >> 
> >> 
> >> Nir
> >> 
> >> On Wed, Aug 17, 2016 at 12:05 PM, Nir Soffer < nsoffer at redhat.com > wrote:
> >> > בתאריך 17 באוג׳ 2016 8:54 לפנה״צ,‏ "Sandro Bonazzola"
> >> > < sbonazzo at redhat.com >
> >> > כתב:
> >> >> 
> >> >> The issue is different:
> >> >> both x86_64 and ppc64le builds python-ioprocess package so a newer
> >> >> python-ioprocess may come from ppc64le build not corresponding to an
> >> >> exact
> >> >> nvr on x86_64.
> >> >> Solution may be to chance the ioprocess build to drop the timestamp or
> >> >> make the ioprocess build a matrix job building from the same src.rpm.
> >> >> Another solution may be relax the nvr requirements in the
> >> >> python-ioprocess
> >> >> package to not require the exact nvr but only the same version.
> >> >> Another solution may be competely separate x86_64 and ppc64le repos.
> >> > 
> >> > I think the issue is the requiring the a version with the timestamp and
> >> > git
> >> > commit, it sould only require the same version.
> >> > 
> >> > 
> >> >> 
> >> >> 
> >> >> 
> >> >> On Sun, Aug 14, 2016 at 8:12 AM, Eyal Edri < eedri at redhat.com > wrote:
> >> >>> 
> >> >>> Was the broken build replaced by new one that is working?
> >> >>> 
> >> >>> On Fri, Aug 12, 2016 at 9:10 PM, Nir Soffer < nsoffer at redhat.com >
> >> >>> wrote:
> >> >>>> 
> >> >>>> Hi all,
> >> >>>> 
> >> >>>> There was a broken build of ioprocess in ovirt repos, causing vdsm
> >> >>>> all
> >> >>>> vdsm master builds to fail with this error:
> >> >>>> 
> >> >>>> ERROR: Command failed. See logs for output.
> >> >>>> 
> >> >>>> If you have failed builds, please retrigger them in jenkins:
> >> >>>> http://jenkins.ovirt.org/gerrit_manual_trigger/
> >> >>>> 
> >> >>>> Nir


-- 
Francesco Romani
RedHat Engineering Virtualization R & D
Phone: 8261328
IRC: fromani