Looking at the load on our hypervisors I'm sure we can add more slaves to
jenkins.
Is there a documented procedure on how to add a new slave so anyone in the
team can do it?
--
Eyal Edri
Associate Manager
RHEV DevOps
EMEA ENG Virtualization R&D
Red Hat Israel
phone: +972-9-7692018
irc: eedri
On 05/10 18:53, Eyal Edri wrote:
> Looking at the load on our hypervisors I'm sure we can add more slaves to
> jenkins.
> Is there a documented procedure on how to add a new slave so anyone in the
> team can do it?
>
I remember writing something about the templates and such, will look for it.
Th
On 05/10 18:02, David Caro wrote:
> On 05/10 18:53, Eyal Edri wrote:
> > Looking at the load on our hypervisors I'm sure we can add more slaves to
> > jenkins.
> > Is there a documented procedure on how to add a new slave so anyone in the
> > team can do it?
> >
>
> I remember writing something a
Hi, I added few more el7 slaves, there is a (relatively) new template
'centos72-jenkins-slave' in the Jenkins_CentOS cluster.
iirc there was a discussion few months ago about our IPs limit, I think
we're approaching that(there are
102 fixed addresses in foreman.phx.ovirt.org dhcpd server, and this
On 05/12 12:46, Nadav Goldin wrote:
> Hi, I added few more el7 slaves, there is a (relatively) new template
> 'centos72-jenkins-slave' in the Jenkins_CentOS cluster.
> iirc there was a discussion few months ago about our IPs limit, I think
> we're approaching that(there are
> 102 fixed addresses in
On 05/12 16:50, Nadav Goldin wrote:
> ok I'll add also FC23, though I think the problem is the vdsm jobs, not sure
> why they use only fc* instead of the el7 ones
Thanks man, that's because (at least the check-merged) use libvirt to start up
vms, and trying to use libvirt from a fc23 chroot on an
ok I'll add also FC23, though I think the problem is the vdsm jobs, not sure
why they use only fc* instead of the el7 ones
On Thu, May 12, 2016 at 4:33 PM, David Caro wrote:
> On 05/12 12:46, Nadav Goldin wrote:
> > Hi, I added few more el7 slaves, there is a (relatively) new template
> > 'cento
sure, added 4 more FC23 slaves, and on the way hit the IPs limit,
had to remove some old unused testing VMs to release IPs.
we have some useless leases there such as an IP for each template,
though I'm not 100% sure if its safe to simply delete them from foreman
(they are saved as templates in the
Please open a ticket in jira on extending the ip range, i guess we need to
handle ir soon
On May 12, 2016 8:58 PM, "Nadav Goldin" wrote:
> sure, added 4 more FC23 slaves, and on the way hit the IPs limit,
> had to remove some old unused testing VMs to release IPs.
> we have some useless leases t
One possibility that will give us a lot of ips is using internal ones, though
it will require a bit more configuration but now that the jenkins master is in
phx I think it might be a good option
David Caro
El 13 may. 2016 7:57, Eyal Edri escribió:
Please open a ticket in jira on extending the
Hello All.
Are those IPs are used by any tests or the jobs itself?
Anton.
On Fri, May 13, 2016 at 9:03 AM, David Caro Estevez
wrote:
> One possibility that will give us a lot of ips is using internal ones,
> though it will require a bit more configuration but now that the jenkins
> master is i
ra" , "dcaro@redhat com" , "Evgheni
Dereveanchin" , "Nadav Goldin"
Sent: Friday, 13 May, 2016 9:03:05 AM
Subject: Re: capacity for more slaves
One possibility that will give us a lot of ips is using internal ones, though
it will require a bit more configuration but no
On 05/13 09:10, Anton Marchukov wrote:
> Hello All.
>
> Are those IPs are used by any tests or the jobs itself?
Are used by the slaves themselves (each slave has one ip), not really needed
for anything (the slaves are not contacted by any outside machine nor offering
any service)
>
> Anton.
>
Hello All.
Then I suggest consider rolling out IPv6 for the slaves right away
especially since it was declared a goal for oVirt 4. This can also simplify
slaves management as we can use SLAC for IP assignment and can solve that
DHCP update problems we had in foreman due to stale mac addresses etc.
On 05/13 10:50, Anton Marchukov wrote:
> Hello All.
>
> Then I suggest consider rolling out IPv6 for the slaves right away
> especially since it was declared a goal for oVirt 4. This can also simplify
> slaves management as we can use SLAC for IP assignment and can solve that
> DHCP update problem
Hello David.
Forgot one "A". So it is SLAAC a scheme where IP address of the host in
IPv6 is generated from the mac address by the host itself and so it is
stateless at that sense and does not require keeping IP to MAC database
anywhere. Also unless privacy extension is enabled the machine will ge
Please add it to the agenda on tomorrow meeting for infra.
e.
On Fri, May 13, 2016 at 2:04 PM, Anton Marchukov
wrote:
> Hello David.
>
> Forgot one "A". So it is SLAAC a scheme where IP address of the host in
> IPv6 is generated from the mac address by the host itself and so it is
> stateless a
17 matches
Mail list logo
