Monday, 2020-11-16

*** zzzeek has quit IRC00:03
*** zzzeek has joined #openstack-infra00:05
*** lifeless has quit IRC00:26
*** ociuhandu has quit IRC00:27
*** lifeless has joined #openstack-infra00:28
*** zzzeek has quit IRC00:28
*** rcernin has quit IRC00:29
*** zzzeek has joined #openstack-infra00:30
*** zxiiro has quit IRC00:30
*** rcernin has joined #openstack-infra00:32
*** sorrison has joined #openstack-infra00:34
*** lifeless has quit IRC00:37
*** ociuhandu has joined #openstack-infra00:41
*** ociuhandu has quit IRC00:47
*** rcernin has quit IRC00:47
*** Goneri has quit IRC00:48
*** lifeless has joined #openstack-infra00:48
*** ociuhandu has joined #openstack-infra00:59
*** ociuhandu has quit IRC01:05
*** ociuhandu has joined #openstack-infra01:18
*** prometheanfire has joined #openstack-infra01:45
*** rcernin has joined #openstack-infra02:04
*** iurygregory has quit IRC02:07
*** takamatsu has quit IRC02:10
*** rcernin has quit IRC02:12
*** takamatsu has joined #openstack-infra02:43
*** dmsimard has quit IRC02:49
*** mmethot_ has joined #openstack-infra02:50
*** dmsimard has joined #openstack-infra02:50
*** mmethot has quit IRC02:53
*** rcernin has joined #openstack-infra03:17
*** zzzeek has quit IRC03:45
*** zzzeek has joined #openstack-infra03:47
*** rcernin has quit IRC03:57
*** reed has quit IRC04:00
*** reed has joined #openstack-infra04:01
*** zzzeek has quit IRC04:20
*** zzzeek has joined #openstack-infra04:20
*** reed has quit IRC04:46
*** reed has joined #openstack-infra04:46
*** rcernin has joined #openstack-infra05:15
*** rcernin has quit IRC05:21
*** evrardjp has quit IRC05:33
*** evrardjp has joined #openstack-infra05:33
*** dviroel has quit IRC05:40
*** rpioso has quit IRC05:40
*** rpioso has joined #openstack-infra05:40
*** dviroel has joined #openstack-infra05:40
*** rcernin has joined #openstack-infra05:47
*** matt_kosut has joined #openstack-infra05:50
*** rcernin has quit IRC05:56
*** reed has quit IRC05:56
*** reed has joined #openstack-infra05:57
*** ianychoi_ has joined #openstack-infra05:57
*** ianychoi has quit IRC06:00
*** rcernin has joined #openstack-infra06:26
*** ricolin has quit IRC06:31
*** rcernin has quit IRC06:31
*** rcernin has joined #openstack-infra06:36
*** lmiccini has joined #openstack-infra06:44
*** reed has quit IRC07:01
*** reed has joined #openstack-infra07:02
*** rcernin has quit IRC07:05
*** sboyron has joined #openstack-infra07:16
*** ricolin_ has joined #openstack-infra07:18
*** ricolin_ has quit IRC07:20
*** ricolin has joined #openstack-infra07:20
*** reed has quit IRC07:22
*** reed has joined #openstack-infra07:22
*** slaweq has joined #openstack-infra07:25
*** rpittau|afk is now known as rpittau07:26
*** iurygregory has joined #openstack-infra07:29
*** rcernin has joined #openstack-infra07:31
*** ralonsoh has joined #openstack-infra07:33
*** rcernin has quit IRC07:36
*** sboyron has quit IRC07:40
*** sshnaidm|off is now known as sshnaidm07:46
*** rcernin has joined #openstack-infra07:50
*** rpittau is now known as rpittau|bbl07:52
*** eolivare has joined #openstack-infra07:55
*** jcapitao has joined #openstack-infra07:55
*** rcernin has quit IRC07:55
*** rcernin has joined #openstack-infra07:56
*** zzzeek has quit IRC08:00
*** zzzeek has joined #openstack-infra08:03
*** hashar has joined #openstack-infra08:10
*** rcernin has quit IRC08:10
*** matt_kosut has quit IRC08:11
*** sboyron has joined #openstack-infra08:11
*** matt_kosut has joined #openstack-infra08:12
*** rcernin has joined #openstack-infra08:13
*** gfidente has joined #openstack-infra08:14
*** reed has quit IRC08:22
*** reed has joined #openstack-infra08:23
*** dchen has quit IRC08:24
*** xek has joined #openstack-infra08:25
*** dchen has joined #openstack-infra08:25
*** dchen has quit IRC08:26
*** rcernin has quit IRC08:37
*** reed has quit IRC08:43
*** reed has joined #openstack-infra08:43
*** ociuhandu has quit IRC08:46
*** andrewbonney has joined #openstack-infra08:49
*** tosky has joined #openstack-infra08:49
*** jpena|off is now known as jpena08:55
*** lucasagomes has joined #openstack-infra08:55
*** zzzeek has quit IRC08:57
*** zzzeek has joined #openstack-infra08:58
*** rcernin has joined #openstack-infra09:04
*** xek has quit IRC09:05
*** zzzeek has quit IRC09:06
*** zzzeek has joined #openstack-infra09:07
*** xek has joined #openstack-infra09:11
*** rcernin has quit IRC09:12
*** zzzeek has quit IRC09:24
*** zzzeek has joined #openstack-infra09:24
*** reed has quit IRC09:28
*** reed has joined #openstack-infra09:28
jpenahi! I see the centos 7 mirror content on the AFS mirrors is outdated, and does not contain the 7.9 rpms09:48
jpenalooking at https://opendev.org/opendev/system-config/src/branch/master/playbooks/roles/mirror-update/files/centos-mirror-update#L51, it looks like we're syncing from mirros.lstn.net09:48
jpenaand it is out of sync, according to https://mirror-status.centos.org/09:48
*** reed has quit IRC09:48
jpenaare we using that server for a special reason, or could we switch to a different one?09:49
*** rcernin has joined #openstack-infra09:49
*** reed has joined #openstack-infra09:49
*** xek has quit IRC09:52
*** xek has joined #openstack-infra09:55
*** rcernin has quit IRC10:03
*** reed has quit IRC10:04
*** reed has joined #openstack-infra10:05
*** sean-k-mooney2 has joined #openstack-infra10:05
*** sean-k-mooney1 has quit IRC10:08
*** reed has quit IRC10:10
*** reed has joined #openstack-infra10:10
*** xek has quit IRC10:11
fricklerjpena: iiuc we keep chasing for the least broken mirror server. if you know a better one, feel free to propose a patch10:13
*** ociuhandu has joined #openstack-infra10:15
*** reed has quit IRC10:20
*** reed has joined #openstack-infra10:20
*** zzzeek has quit IRC10:23
*** zzzeek has joined #openstack-infra10:24
*** reed has quit IRC10:25
*** reed has joined #openstack-infra10:26
*** sean-k-mooney has joined #openstack-infra10:27
*** sean-k-mooney2 has quit IRC10:30
*** dtantsur|afk is now known as dtantsur10:30
*** sean-k-mooney1 has joined #openstack-infra10:32
*** sean-k-mooney has quit IRC10:35
*** zzzeek has quit IRC10:41
*** zzzeek has joined #openstack-infra10:42
*** reed has quit IRC10:46
*** reed has joined #openstack-infra10:46
openstackgerritLiam Young proposed openstack/project-config master: Add Magpie charm to OpenStack charms  https://review.opendev.org/76281910:59
*** reed has quit IRC11:06
*** reed has joined #openstack-infra11:06
*** zzzeek has quit IRC11:12
*** zzzeek has joined #openstack-infra11:13
*** reed has quit IRC11:21
*** reed has joined #openstack-infra11:22
*** reed has quit IRC11:27
*** zzzeek has quit IRC11:27
*** sean-k-mooney1 has quit IRC11:27
*** reed has joined #openstack-infra11:27
*** sean-k-mooney1 has joined #openstack-infra11:27
*** sean-k-mooney1 is now known as sean-k-mooney11:28
*** rpittau|bbl is now known as rpittau11:30
*** zzzeek has joined #openstack-infra11:30
*** rcernin has joined #openstack-infra11:30
*** reed has quit IRC11:32
*** reed has joined #openstack-infra11:32
*** reed has quit IRC11:37
*** reed has joined #openstack-infra11:38
*** ociuhandu has quit IRC11:42
*** reed has quit IRC11:48
*** reed has joined #openstack-infra11:48
*** jcapitao is now known as jcapitao_lunch11:52
*** rcernin has quit IRC11:56
*** thiagop has joined #openstack-infra11:58
*** lpetrut has joined #openstack-infra12:00
*** rcernin has joined #openstack-infra12:02
*** ociuhandu has joined #openstack-infra12:13
*** zzzeek has quit IRC12:14
*** zzzeek has joined #openstack-infra12:16
*** rcernin has quit IRC12:19
*** rfolco has joined #openstack-infra12:19
*** jpena is now known as jpena|lunch12:31
*** rlandy has joined #openstack-infra12:33
*** rlandy is now known as rlandy|rover12:33
*** xek has joined #openstack-infra12:35
*** xek_ has joined #openstack-infra12:38
*** xek has quit IRC12:41
*** raissa has joined #openstack-infra12:48
*** jcapitao_lunch is now known as jcapitao12:50
*** raissa has quit IRC12:53
*** reed has quit IRC13:08
*** reed has joined #openstack-infra13:08
*** reed has quit IRC13:18
*** reed has joined #openstack-infra13:19
*** ociuhandu has quit IRC13:19
*** weshay|ruck has joined #openstack-infra13:22
*** ociuhandu has joined #openstack-infra13:22
*** Goneri has joined #openstack-infra13:27
*** jpena|lunch is now known as jpena13:29
*** zzzeek has quit IRC13:31
*** zzzeek has joined #openstack-infra13:34
*** d34dh0r53 has joined #openstack-infra13:45
*** slaweq has quit IRC13:47
*** tbarron has joined #openstack-infra13:47
*** slaweq has joined #openstack-infra13:50
weshay|ruckianw, have you seen https://review.opendev.org/#/c/76281213:52
weshay|ruckzbr, hey13:54
zbrweshay|ruck: hi13:54
*** zzzeek has quit IRC14:02
*** zzzeek has joined #openstack-infra14:07
*** zzzeek has quit IRC14:14
*** zzzeek has joined #openstack-infra14:14
*** zxiiro has joined #openstack-infra14:21
*** reed has quit IRC14:29
*** reed has joined #openstack-infra14:29
*** dtroyer has quit IRC14:31
zbrfungi: clarkb: around to help with the mirroring swap?14:33
*** zzzeek has quit IRC14:33
*** zzzeek has joined #openstack-infra14:36
fungidid someone find a more current rsync mirror to switch to? every time we pick one, it eventually breaks or goes stale and we have to pick another14:39
*** artom has joined #openstack-infra14:40
*** zzzeek has quit IRC14:41
*** zzzeek has joined #openstack-infra14:45
zbrfungi: that is why CDNs were invented.14:50
zbrthey are supposed to resolve the burden of maintaining list of mirrors,... if they work fine.14:51
*** xek__ has joined #openstack-infra14:52
*** reed has quit IRC14:54
*** reed has joined #openstack-infra14:55
*** xek_ has quit IRC14:55
*** beekneemech is now known as bnemec15:00
*** reed has quit IRC15:05
*** reed has joined #openstack-infra15:05
*** lpetrut has quit IRC15:05
*** reed has quit IRC15:10
*** reed has joined #openstack-infra15:10
*** zzzeek has quit IRC15:11
*** zzzeek has joined #openstack-infra15:14
*** ociuhandu has quit IRC15:32
*** hashar has quit IRC15:32
*** ociuhandu has joined #openstack-infra15:32
*** ociuhandu has quit IRC15:33
*** ociuhandu has joined #openstack-infra15:33
*** dave-mccowan has joined #openstack-infra15:34
*** redrobot has quit IRC15:35
*** xek_ has joined #openstack-infra15:37
*** xek__ has quit IRC15:40
*** dave-mccowan has quit IRC15:45
*** hashar has joined #openstack-infra15:48
*** zzzeek has quit IRC15:49
*** zzzeek has joined #openstack-infra15:50
*** Guest64097 has joined #openstack-infra15:51
*** Guest64097 is now known as redrobot15:53
fungizbr: yes, point me to the rsync cdn then?15:55
fungii'm sure it will solve our problems once and for all15:55
fungilogan-: if you're around, not sure if you have any insight into the mirror.lstn.net service but it's apparently got stale centos 7.x packages15:56
zbrlogan was idea for the last 5 days.15:57
zbridle15:57
zbrlogan was idle15:57
zbrfungi: i am trying to understand how https://review.opendev.org/#/c/762812/ can make the mirroring worse than now.15:58
*** zzzeek has quit IRC15:59
clarkbzbr: rsync tries to connect and fails every time it attempts to connect because of connection limits16:00
clarkbthen you end up stale because you're never syncing16:00
*** zzzeek has joined #openstack-infra16:00
clarkbright now we'll get updated if our upstream updates16:00
clarkbthat is better16:00
fungizbr: also i was unaware of 762812 until just now. i suppose that was some missing context from your question above16:03
*** dklyle has joined #openstack-infra16:05
zbryep that is the change. I am not sure about the mentioned connection limits, maybe someone can explain them to me.16:06
fungii'm not opposed to going back to trying mirrors.k.o, but as mentioned by others our past experience is that most of the time we try to connect we get an error saying the connection limit has been reached so we rarely manage to pull updates from there16:07
clarkbzbr: the kernel.org rsync mirrors limit the number of rsync connections to it. They are a very popular mirror so are basically always at the limit. When they are at the limit our rsync connections to them fail and we don't sync16:07
fungiespecially if there's a major centos or similar update, we've seen it go weeks without an open slot for connections16:07
clarkbat least with the current situation when the current mirror does update we will update. If we switch to kernel.org we likely won't update reliably ever16:08
zbrhow how load do these limits stand for?16:08
fungiand we try to be polite (probably unlike others pulling from there) so we don't aggressively reconnect we just try once every couple hours16:08
zbrare they implemented at DNS level or at specific edge server?16:08
clarkbyou cannot implement connection limits in dns16:08
fungizbr: they are global. x number of clients allowed to connect to the mirror16:08
clarkbits tcp level go away16:09
fungiso something like if 25 people are downloading from it at the same time it serves up an error to anyone else who connects until the connection count drops below their limit16:09
zbrok but I do not expect it to continue to fail for a very long period of time.16:09
fungi(i don't recall the exact number)16:09
clarkbI think it was 5016:10
fungilike i've said, we've seen it take weeks to pull an update from there in the past. maybe it's better now? i'm not opposed to trying again, just setting expectations16:10
clarkbwhy not use a different mirror that we don't expect to fail?16:10
*** reed has quit IRC16:10
*** reed has joined #openstack-infra16:11
clarkbI think that was the point of pabelanger's comment16:11
clarkb"kernel.org is expected to fail, let's find another alternative"16:11
fungithe counterargument is that other mirrors will eventually run into problems and cease updating for extended periods of time, while mirrors.k.o has someone grooming the round robin list and removing stale mirrors16:11
zbrfungi: yep, that was my point. that someone else is supposed to maintain the list of actively maintained mirror.16:12
*** jamesdenton has quit IRC16:13
fungiwell, not exactly round-robin, looks like it relies on dns-based load distribution like 3dns or similar16:13
* zbr is also aware that whatever we do is a gamble, but lets try to make it an informed gamble.16:13
*** jamesdenton has joined #openstack-infra16:13
clarkbinteresting that centos mirror list has kernel.org mirrors in california (they are in oregon)16:15
fungifor me it resolves to something that claims to be in/near san francisco16:16
clarkbnot that important, and ya if we want to try it and see if it works then we can do that. But it would also probably be a good idea to look for other alternatives as well16:16
clarkbfungi: huh, I checked from my ovh node and home and they both point to oregon. I guess they don't exclusively host with osuosl then16:16
clarkbzbr: to be clear I'm not sure you assertion in the change is correct. We have seen it break perpetually for long periods in the past16:17
*** zzzeek has quit IRC16:17
clarkbits more like you succeeds once every few weeks if at all vs fail once every few weeks if at all16:17
*** zzzeek has joined #openstack-infra16:20
fungilooks like the last time we tried it for centos may have been before https://review.openstack.org/551318 switched to limestone a couple years ago16:20
fungithough we've more recently had to stop mirroring epel from there or opensuse for various reasons16:20
zbralmost 3 years. i guess reverting the change could allow us to see if situation improved.16:22
zbrif it does not, we can start benchmarking specific mirrors16:22
*** reed has quit IRC16:23
fungihttps://review.openstack.org/553052 is when we moved off mirrors.k.o for epel, and was also around the same time16:24
fungialso cites "connection issues" so i assume that was the max clients thing as well16:25
fungia year ago we moved opensuse mirroring around because mirrors.k.o and others had outdated content16:25
clarkbthe big issue with rsync connection limits aiui is that connections tend to be held open for long periods of time when they are used16:25
clarkbwhich causes them to fill the pool quickly.16:25
clarkbthe way around that is to authenticate and bypass the pool limits for anonymous access, but to do that you typically need to be a proper first level mirror and we're not interested in doing that as the requirements are much more stringent16:26
clarkbbasically we want to mirror for our CI system not for everyone16:26
clarkb(fwiw I THink that is a reasonable requirement on the part of the mirroring systems out there. They have limited bandwidth so distributing to as wide an audience as possible first is a good idea)16:30
zbrclarkb: do you know were is the rsync mirroring documented? I was not able to find any link on hints on http://mirrors.kernel.org/16:39
clarkbzbr: I don't. I know the rsync urls are listed on the mirror list though16:39
zbri do see disclaimer but not a clue about how is setup, limitations,... eventually the tiers.16:40
*** eolivare has quit IRC16:44
*** reed has joined #openstack-infra16:49
*** eolivare has joined #openstack-infra16:53
*** mgoddard has joined #openstack-infra16:55
*** lucasagomes has quit IRC17:03
*** amoralej|off has joined #openstack-infra17:03
*** amoralej|off has left #openstack-infra17:03
fungijpena: weshay|ruck: zbr: the mirror change has landed, it should hopefully be in place by the time the next rsync is attempted at 18:43 utc. you'll be able to see the logs at https://static.opendev.org/mirror/logs/rsync-mirrors/centos.log17:09
zbrthanks17:09
weshay|ruckfungi++ clarkb++17:10
weshay|ruckthanks17:10
weshay|ruckrlandy|rover, ^^17:10
*** hamalq has joined #openstack-infra17:12
rlandy|rovergreat - thanks17:12
*** jcapitao has quit IRC17:14
*** kwazar has joined #openstack-infra17:35
*** zzzeek has quit IRC17:37
*** zzzeek has joined #openstack-infra17:38
*** ociuhandu_ has joined #openstack-infra17:38
*** rpittau is now known as rpittau|afk17:41
*** ociuhandu has quit IRC17:41
*** ociuhandu_ has quit IRC17:43
*** zzzeek has quit IRC17:45
*** zzzeek has joined #openstack-infra17:48
*** ociuhandu has joined #openstack-infra17:49
*** ociuhandu has quit IRC17:54
*** mgoddard has quit IRC18:05
*** jpena is now known as jpena|off18:10
faruqhi could really really use some help - I am trying to install charmed openstack on three KVM from maas and I keep getting hung on the mysql-innodb-cluster. It tries and then it just gets stuck18:22
clarkbfaruqh: we offer help with the developer infrastructure used to build the openstack software here. We don't really do much with the deployment side of openstack itself18:23
clarkbfaruqh: I think there is an openstack charms channel though which might be a better place for this question. Let me see if I can find it18:23
faruqhiawesome clarkb thanks18:23
clarkbhrm no they have held meetings int he past but no separate channel that I see18:24
clarkbstill looking though18:24
faruqhthank you -18:24
clarkbah it may be #openstack-charms according to their docs18:24
faruqhclarkbthanks alot18:25
clarkbyou're welcome18:25
*** gfidente is now known as gfidente|afk18:37
*** eolivare has quit IRC18:53
*** zzzeek has quit IRC18:59
*** reed has quit IRC18:59
*** reed has joined #openstack-infra18:59
*** zzzeek has joined #openstack-infra19:00
*** andrewbonney has quit IRC19:10
fungijpena: weshay|ruck: zbr: rlandy|rover: the centos 7.x mirror has updated, see the end of the log for what was updated... https://static.opendev.org/mirror/logs/rsync-mirrors/centos.log19:30
* weshay|ruck looks.. thanks fungi 19:30
rlandy|roverfungi: thank - job running to test centos-7 now19:31
fungiwe'll want to keep an eye on it to make sure we don't start running into the rsync server being full on its client limit19:39
fungisince we've seen m.k.o do that for extended periods in the past19:40
*** dciabrin has quit IRC19:52
*** dciabrin has joined #openstack-infra19:52
*** reed has quit IRC19:54
*** reed has joined #openstack-infra19:55
*** reed has quit IRC19:59
*** reed has joined #openstack-infra19:59
*** thiagop has quit IRC20:00
*** ralonsoh has quit IRC20:00
*** dtantsur is now known as dtantsur|afk20:02
*** reed_ has joined #openstack-infra20:11
*** reed has quit IRC20:12
*** reed has joined #openstack-infra20:12
*** faruqh has quit IRC20:14
*** hashar has quit IRC20:21
*** iurygregory has quit IRC20:40
*** iurygregory has joined #openstack-infra20:43
*** raissa has joined #openstack-infra20:44
*** ociuhandu has joined #openstack-infra20:55
*** ociuhandu has quit IRC20:57
*** arne_wiebalck has quit IRC21:13
*** arne_wiebalck has joined #openstack-infra21:13
*** lbragstad_ has joined #openstack-infra21:21
*** lbragstad has quit IRC21:24
*** rcernin has joined #openstack-infra21:50
*** rcernin has quit IRC21:51
*** rcernin has joined #openstack-infra21:51
*** matt_kosut has quit IRC21:57
*** rcernin has quit IRC22:02
*** matt_kosut has joined #openstack-infra22:04
*** rcernin has joined #openstack-infra22:06
*** matt_kosut has quit IRC22:08
*** sboyron has quit IRC22:08
*** sboyron has joined #openstack-infra22:09
*** rlandy|rover is now known as rlandy|rover|bbl22:28
*** xek_ has quit IRC22:29
*** lbragstad__ has joined #openstack-infra22:35
*** rledisez has quit IRC22:36
melwittI'm looking at an example of a TIMED_OUT job and wondering why don't I see the devstack and tempest steps (and their timings) in the console? I feel like they used to be there https://zuul.openstack.org/build/75d7b93f5385449589540c2657422841/console22:36
*** rledisez has joined #openstack-infra22:36
clarkbmelwitt: it is likely due to the fact that when a job is timed out ansible is forceibly killed. And when that happens the json object that console renders from is not properly finished up22:36
clarkbmelwitt: in those cases the console log should hae the info you need /me looks too22:37
melwittah ok. I can look at timings in the job-output.txt but it would be easier if it were in the console. just trying to identify any trends in why certain jobs time out more than others22:38
*** lbragstad_ has quit IRC22:38
clarkbya the reason it isn't there is ansible was killed during the run phase due to a timeout22:39
clarkband that means that we don't write out a properly formatted json file for the console to render22:39
melwittI see22:39
*** slaweq has quit IRC22:39
fungii wonder if we could do a better job of redirecting folks to the console log in those situations22:40
clarkbhttps://f75b41bb61072b844b31-b8d0e37c8407321b7ba0b13922acea87.ssl.cf1.rackcdn.com/735570/2/check/nova-ceph-multistore/75d7b93/job-output.json shows this as it goes pre pre pre post22:40
melwitthm ok22:40
melwittI could have sworn the devstack step was separate from the tempest step, but it must all fall under "run" I guess22:43
clarkbyes in those jobs I think they are both part of the run playbook22:43
melwittgotcha, I understand now22:43
clarkbsome jobs put devstack in pre because they are less concerned with whether or not a working cloud is produced and just want to assume one exists (sdk tests for example)22:43
melwittI see22:45
*** slaweq has joined #openstack-infra22:48
*** tosky_ has joined #openstack-infra22:52
*** tosky is now known as Guest4141922:55
*** tosky_ is now known as tosky22:55
*** Guest41419 has quit IRC22:55
*** slaweq has quit IRC22:58
*** sboyron has quit IRC23:03
*** slaweq has joined #openstack-infra23:07
*** slaweq has quit IRC23:13
*** gibi has quit IRC23:20
*** gibi has joined #openstack-infra23:27
*** dchen has joined #openstack-infra23:41
*** raissa has quit IRC23:48

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