Re: Rescheduled meeting
- Original Message - > From: "R P Herrold" > To: "Ewoud Kohl van Wijngaarden" > Cc: "oVirt infrastructure ML" > Sent: Monday, December 2, 2013 8:11:30 PM > Subject: Rescheduled meeting > > On Mon, 2 Dec 2013, Ewoud Kohl van Wijngaarden wrote: > > > I'd propose we do it by mail and we'll meet again next week. Please add > > where you can. > > > * update a ticket on rackspace and raise its priority > > I escalated this last week at oVirt Weekly Sync > https://fedorahosted.org/ovirt/ticket/80 > > dcaro, dneary? i think we should request an alternate DRAC/KVM interface from RACKSPACE, this ticket/issue has been open for way too long, and mostly because of thier inability to provide us console access to the server. > > > * We should find a solution to do regular yum updates. This can be as > > simple as a weekly or monthly manual check where we do updates and > > reboot if needed. Anyone knows a good solution? > > The problem is that to be safe, one wants a testing, dev, prod > approach to test in and to permit loss of services on > production. Until the '03 at Rackspace comes on line we > cannot do that > > -- Russ herrold > ___ > 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
Re: Rescheduled meeting
- Original Message - > From: "Ewoud Kohl van Wijngaarden" > To: infra@ovirt.org > Sent: Monday, December 2, 2013 5:17:55 PM > Subject: Rescheduled meeting > > Due to a scheduling conflict there was also another IRC meeting planned > in our regular timeslot. I offered to reschedule ours, but I don't know > what another time slot would be. > > I'd propose we do it by mail and we'll meet again next week. Please add > where you can. > > = Action items from last week = > > * update a ticket on rackspace and raise its priority > > * ewoud rename the puppet branch > Haven't worked on it. Will most likely need to work with dcaro because I > don't > have the permissions. > > * work on ticket to create a jenkins job to automate the deployment after a > merge on puppet > This is still blocked on the puppet branch rename. > > * eedri to merge http://gerrit.ovirt.org/#/c/21094/ > Was merged > > * upgrade/reinstall all f18 slaves to f20 > Some F20 slaves were added. Eedri? we already had one baremetal host running f20, and i also upgraded jenkins-slave-vm02 to fedora20. i think we might need more slaves soon, maybe worth upgrading current f18 hosts. > > = Hosting = > > = Puppet and Foreman = > > * Foreman was upgraded to 1.3. I forgot to formally announce it and > downtime was longer than it should have been. Ran into > http://projects.puppetlabs.com/issues/23315 (puppetdb + java 1.6 + > openssl from EL6.5 aren't friends). Solved by installing java 1.7. > > * With Foreman 1.3 we can associate VMs to compute resources. This means > our VMs that were installed and later added to Foreman now also have > the compute resource integation. That means console and power > controls. There are still 2 machines not associated (jenkins vm01 and > vm03) and I don't know why yet. > > = Jenkins= i had to restart jenkins after seeing this issue - http://paste.fedoraproject.org/58517/60206531/ i also see lots of exceptions and errors on the log. we need to allocate time to review those errors and do a scan on existing plugins - what do we need, what can be updated/removed? we need to upgrade our existing CENTOS 6.4 TO CENTOS 6.5 which was just released. any ideas when/how? > > = Other business? = > > * Minutes URLs on http://www.ovirt.org/Infrastructure_team_meetings point > to the logs instead of the nicer minutes. Compare: > http://resources.ovirt.org/meetings/ovirt/2013/ovirt.2013-11-25-15.05.html > > http://resources.ovirt.org/meetings/ovirt/2013/ovirt.2013-11-25-15.05.log.html > I'd propose we link the short version since that includes a link to > the full logs, but not the other way around. > > * We should find a solution to do regular yum updates. This can be as > simple as a weekly or monthly manual check where we do updates and > reboot if needed. Anyone knows a good solution? agree, plus regular jenkins / other infra servers upgrades / maintenance. > ___ > 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
Re: Rescheduled meeting
On Mon, Dec 02, 2013 at 04:17:55PM +0100, Ewoud Kohl van Wijngaarden wrote: > * With Foreman 1.3 we can associate VMs to compute resources. This means > our VMs that were installed and later added to Foreman now also have > the compute resource integation. That means console and power > controls. There are still 2 machines not associated (jenkins vm01 and > vm03) and I don't know why yet. After some investigation I found that the facter value for macaddress was returning the value for the vdsm dummy interface. Manual override works: First ensure foreman knows the host by the correct MAC address at least once: FACTER_MACADDRESS=00:1a:4a:1e:08:41 puppet agent -t Then go to the compute resource in foreman, associate VMs and it should be able to accociate it. ___ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
Rescheduled meeting
On Mon, 2 Dec 2013, Ewoud Kohl van Wijngaarden wrote: > I'd propose we do it by mail and we'll meet again next week. Please add > where you can. > * update a ticket on rackspace and raise its priority I escalated this last week at oVirt Weekly Sync https://fedorahosted.org/ovirt/ticket/80 dcaro, dneary? > * We should find a solution to do regular yum updates. This can be as > simple as a weekly or monthly manual check where we do updates and > reboot if needed. Anyone knows a good solution? The problem is that to be safe, one wants a testing, dev, prod approach to test in and to permit loss of services on production. Until the '03 at Rackspace comes on line we cannot do that -- Russ herrold ___ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
Rescheduled meeting
Due to a scheduling conflict there was also another IRC meeting planned in our regular timeslot. I offered to reschedule ours, but I don't know what another time slot would be. I'd propose we do it by mail and we'll meet again next week. Please add where you can. = Action items from last week = * update a ticket on rackspace and raise its priority * ewoud rename the puppet branch Haven't worked on it. Will most likely need to work with dcaro because I don't have the permissions. * work on ticket to create a jenkins job to automate the deployment after a merge on puppet This is still blocked on the puppet branch rename. * eedri to merge http://gerrit.ovirt.org/#/c/21094/ Was merged * upgrade/reinstall all f18 slaves to f20 Some F20 slaves were added. Eedri? = Hosting = = Puppet and Foreman = * Foreman was upgraded to 1.3. I forgot to formally announce it and downtime was longer than it should have been. Ran into http://projects.puppetlabs.com/issues/23315 (puppetdb + java 1.6 + openssl from EL6.5 aren't friends). Solved by installing java 1.7. * With Foreman 1.3 we can associate VMs to compute resources. This means our VMs that were installed and later added to Foreman now also have the compute resource integation. That means console and power controls. There are still 2 machines not associated (jenkins vm01 and vm03) and I don't know why yet. = Jenkins= = Other business? = * Minutes URLs on http://www.ovirt.org/Infrastructure_team_meetings point to the logs instead of the nicer minutes. Compare: http://resources.ovirt.org/meetings/ovirt/2013/ovirt.2013-11-25-15.05.html http://resources.ovirt.org/meetings/ovirt/2013/ovirt.2013-11-25-15.05.log.html I'd propose we link the short version since that includes a link to the full logs, but not the other way around. * We should find a solution to do regular yum updates. This can be as simple as a weekly or monthly manual check where we do updates and reboot if needed. Anyone knows a good solution? ___ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra