Tuesday, 2021-05-11

*** hamalq has quit IRC00:06
*** timburke has joined #opendev00:26
*** artom has quit IRC00:28
*** hemanth_n has joined #opendev00:41
openstackgerritTim Burke proposed opendev/system-config master: Link to www.openstack.org instead of openstack.org  https://review.opendev.org/c/opendev/system-config/+/79053300:42
fungiclarkb: will 789622 need updating to match ^ ?00:47
*** hemanth_n has quit IRC00:50
*** pongboom2 has joined #opendev01:10
timburkeor we fix the openstack.org cert -- i'm happy to abandon ;-)01:13
*** pongboom has quit IRC01:13
fungitimburke: i'll bring it to the attention of the foundation webdevs01:35
fungithanks for flagging. for some reason https://openstack.org/ seems to be serving a cert which expired saturday01:36
*** ysandeep|SL is now known as ysandeep01:54
*** hemanth_n has joined #opendev02:09
*** whoami-rajat has quit IRC02:18
*** ysandeep is now known as ysandeep|afk03:23
*** ykarel has joined #opendev04:53
*** ralonsoh has joined #opendev05:10
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Add fedora-containerfile element  https://review.opendev.org/c/openstack/diskimage-builder/+/79036505:18
openstackgerritIan Wienand proposed openstack/diskimage-builder master: bootloader: disable BLS for Fedora  https://review.opendev.org/c/openstack/diskimage-builder/+/79054905:18
*** fdegir has quit IRC05:26
*** fdegir has joined #opendev05:26
*** mkowalski has quit IRC05:27
*** cgoncalves has quit IRC05:27
*** cgoncalves has joined #opendev05:27
*** mkowalski has joined #opendev05:27
*** prometheanfire has quit IRC05:27
*** dhellmann has quit IRC05:28
*** prometheanfire has joined #opendev05:28
*** dhellmann has joined #opendev05:30
*** amoralej|off is now known as amoralej05:55
*** vishalmanchanda has joined #opendev06:22
*** ysandeep|afk is now known as ysandeep06:48
openstackgerritIan Wienand proposed openstack/diskimage-builder master: bootloader: disable BLS for Fedora  https://review.opendev.org/c/openstack/diskimage-builder/+/79054906:49
*** sboyron has joined #opendev06:49
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Add fedora-containerfile element  https://review.opendev.org/c/openstack/diskimage-builder/+/79036506:49
*** fressi has joined #opendev07:01
*** sshnaidm|afk has quit IRC07:03
*** mkowalski has quit IRC07:08
*** mkowalski has joined #opendev07:09
*** marios has joined #opendev07:13
*** sshnaidm has joined #opendev07:20
*** hashar has joined #opendev07:22
*** andrewbonney has joined #opendev07:23
*** mkowalski has quit IRC07:26
*** rpittau|afk is now known as rpittau07:38
*** tosky has joined #opendev07:39
*** mgoddard has quit IRC07:45
*** jpena|off is now known as jpena07:51
*** marios has quit IRC08:18
*** lpetrut has joined #opendev08:22
*** marios has joined #opendev08:31
openstackgerritLajos Katona proposed openstack/project-config master: Rename x/tap-as-a-service to openstack/tap-as-a-service  https://review.opendev.org/c/openstack/project-config/+/79009308:43
*** mkowalski has joined #opendev09:05
*** Kvisle has joined #opendev09:05
KvisleHi! Quick question - looking at https://docs.opendev.org/opendev/system-config/latest/contribute-cloud.html, "Since there’s a bit of setup and maintenance involved in adding a new provider, a minimum of 100 instances would be helpful." ... is this correct, or is smaller amounts than this of interest? (The cloud mostly provides instances with Intel Gold 6126/6226/6226R CPUs in a no-overcommit setup, with09:11
Kvislelocal ssds and a lot more memory than requried) ... but 100 instances would mean booking 12 physical servers -- which exceeds the amount we have to spare.09:11
*** ykarel_ has joined #opendev09:19
*** ykarel has quit IRC09:21
*** ykarel_ is now known as ykarel09:28
ianwKvisle: we can certainly be more flexible09:41
ianwnote at any time you can see the current providers, and their configuration @ https://opendev.org/openstack/project-config/src/branch/master/nodepool09:42
*** ykarel is now known as ykarel|lunch09:54
KvisleWould a contribution be of value, or do you feel that you have more capacity than you need?09:57
*** ykarel_ has joined #opendev09:58
Kvisle(It's not obvious for me from looking in git + zuul)09:58
*** ykarel|lunch has quit IRC10:00
*** dtantsur|afk is now known as dtantsur10:15
ianwKvisle: you can see our node usage @ https://grafana.opendev.org/d/KRmop6EMz/nodepool?orgId=110:24
ianwand you can see Zuul queues @ https://grafana.opendev.org/d/5Imot6EMk/zuul-status?orgId=110:25
*** lpetrut has quit IRC10:30
ianwi think we would definitely be interested in talking more.  clarkb is probably the best person, and will be around in US time10:30
*** ykarel_ is now known as ykarel10:56
*** lpetrut has joined #opendev10:56
*** ysandeep is now known as ysandeep|afk11:11
*** jpena is now known as jpena|lunch11:15
*** lpetrut has quit IRC11:27
*** ysandeep|afk is now known as ysandeep11:40
*** jpena|lunch is now known as jpena11:56
*** dpawlik has quit IRC12:12
*** ykarel_ has joined #opendev12:18
*** dpawlik3 has joined #opendev12:20
*** ykarel has quit IRC12:20
*** amoralej is now known as amoralej|lunch12:22
*** dpawlik3 is now known as dpawlik12:26
Kvisleianw: interesting - I'll see if I can get some contact tonight, then12:45
Kvisleianw: our cloud is in the nordics, so some latency are to be expected - but that should probably not be a big deal?12:45
*** jaicaa has quit IRC12:47
*** jaicaa has joined #opendev12:47
fricklerKvisle: we also use resources from ovh in france, so that should be fine. we set up one mirror host within each cloud, too, so that most of the traffic should be local12:53
fricklernot sure how the ratio would be, though, or how our traffic requirements in general look like12:54
fricklerKvisle: but I agree with ianw, 100 instances for sure isn't a hard limit. and I'm pretty sure we wouldn't claim to have too much capacity available12:56
*** hemanth_n has quit IRC13:06
*** amoralej|lunch is now known as amoralej13:24
parallaxHi - my company's considering the same - we're willing to contribute some of the capacity to OpenDev  - struggling with the IPV6 access currently though;  once the setup's done, around 512 cores could be provided13:24
*** marios is now known as marios|call13:29
*** artom has joined #opendev13:41
*** ykarel_ has quit IRC13:42
*** ykarel_ has joined #opendev13:44
*** stephenfin has quit IRC13:56
*** marios|call is now known as marios13:58
*** mlavalle has joined #opendev13:58
*** fressi has quit IRC14:18
*** iurygregory has quit IRC14:22
*** iurygregory has joined #opendev14:22
*** ykarel_ is now known as ykarel14:25
*** hashar is now known as hasharAway14:29
*** stephenfin has joined #opendev14:34
*** ykarel has quit IRC14:42
clarkbfungi: yes if we want to aldn that then would need to update. Your comments indicate we don't need to land 790533 anymore though?14:50
clarkbparallax: Kvisle  that document should probably be updated. 100 was sort of a round number to give people a sense for what we are looking for but since then it has become easier to set up a new cloud and we're also finding that things like IP addresses are our major limitation and that value may not be appropriate anymore14:52
clarkbI'm definitely willing to have a conversation with both of you to answer any questions and clarify things14:52
openstackgerritMerged openstack/project-config master: Retire puppet-glare - Step 1: End project Gating  https://review.opendev.org/c/openstack/project-config/+/79005614:56
clarkbThe only issue is today (and possibly tomorrow) are not great for me. I've got errands away from the computer. If you see these messages feel free to start mail threads at service-discuss@lists.opendev.org and we can take it from there14:57
*** amoralej is now known as amoralej|off15:09
parallaxclarkb: Sure - that works. Thanks15:09
*** Green_Bird has joined #opendev15:12
clarkbianw: responded to your comment on the zuul02 stack. Can you take a look and see if you think that change is worthwhile? if so I can make those updates15:34
openstackgerritClark Boylan proposed opendev/system-config master: Ansible mailman configs  https://review.opendev.org/c/opendev/system-config/+/78962215:40
openstackgerritClark Boylan proposed opendev/system-config master: Add infra-prod-service-lists job  https://review.opendev.org/c/opendev/system-config/+/79052415:40
clarkbianw: ^ also I like your jinja much more. Thanks!15:40
*** ysandeep is now known as ysandeep|away15:48
*** ykarel has joined #opendev16:00
*** hasharAway is now known as hashar16:02
*** lbragstad has quit IRC16:04
*** marios is now known as marios|out16:10
*** lbragstad has joined #opendev16:16
*** ralonsoh has quit IRC16:25
*** marios|out has quit IRC16:26
*** ykarel has quit IRC16:29
*** dtantsur is now known as dtantsur|afk16:32
*** sboyron has quit IRC16:34
*** rpittau is now known as rpittau|afk16:36
*** fressi has joined #opendev16:54
*** fressi has quit IRC16:59
*** jpena is now known as jpena|off17:07
*** andrewbonney has quit IRC17:07
mnasiadkaIs there a nice way to specify kernel cmdline args for a nodepool instance other than specifying that in the image configuration in project-config?17:11
clarkbmnasiadka: you may need to describe the problem a bit more. but generally the kernel cmdline is specified in our image builds. Jobs can reboot if necessary as well (which allows you to change stuff like that potentially)17:13
mnasiadkaclarkb: that’s my plan (reboot) - Debian Bullseye disables cgroup v2 by default and Kolla uses libvirt in a container - so it can’t rely on systemd to manage cgroups v217:14
fungiyeah, to restate, if specific jobs need a particular kernel option set at boot, they'll need to reboot in the job to do that17:14
mnasiadkaOk then, thanks for confirmation :)17:15
fungias an example, mnasiadka: you might want to take a look at the work ade is doing for fips: https://review.opendev.org/78877817:16
clarkbI guess the issue is that you cannot have v1 and v2 enabled at the same time. In that case we probably don't want to change the bullseye defaults17:24
*** hashar is now known as hasharAway17:27
fungiright, but also more generally this is working as designed, kolla has a clear reminder that using v2 on bullseye won't work without adjusting the kernel config, so this is presenting a fairly good example of what a user's bullseye install would look like too17:36
clarkb++17:37
*** klonn has joined #opendev17:46
*** klonn has quit IRC17:47
*** lpetrut has joined #opendev17:55
*** lpetrut has quit IRC17:57
corvusclarkb: is 790798 because you're bootstrapping zuul02 without a statsd config?18:02
clarkbcorvus: no, it is due to someone having problems on the zuul discuss list18:02
clarkbcorvus: http://lists.zuul-ci.org/pipermail/zuul-discuss/2021-May/001584.html18:02
corvusah thx18:03
clarkbit shouldn't matter for zuul02 since we will only start it after zuul01 has stopped18:03
*** whoami-rajat_ has joined #opendev18:33
*** vishalmanchanda has quit IRC18:34
*** hamalq has joined #opendev18:50
*** yoctozepto has quit IRC19:13
*** yoctozepto6 has joined #opendev19:13
*** hamalq has quit IRC20:09
*** erbarr has quit IRC20:11
*** mordred has quit IRC20:11
*** hamalq has joined #opendev20:13
*** erbarr has joined #opendev20:13
*** mordred has joined #opendev20:13
*** snbuback has joined #opendev20:15
*** mordred has quit IRC20:16
*** dirk has quit IRC20:17
*** hasharAway has quit IRC20:27
*** mordred has joined #opendev20:41
*** dirk has joined #opendev20:42
*** mordred has quit IRC20:57
*** Green_Bird has quit IRC21:02
*** dirk has quit IRC21:02
*** slaweq has quit IRC21:19
*** mordred has joined #opendev21:26
*** dirk has joined #opendev21:30
ianwclarkb: zuul02 change lgtm as is21:48
*** whoami-rajat_ is now known as whoami-rajat23:05
*** tosky has quit IRC23:24
*** mlavalle has quit IRC23:44

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!