Tuesday, 2019-10-15

*** spatel has joined #openstack-lbaas00:25
openstackgerritMichael Johnson proposed openstack/octavia master: Fix issues with unavailable secrets  https://review.opendev.org/68822100:28
*** tkajinam_ has quit IRC00:49
*** tkajinam has joined #openstack-lbaas00:49
*** AlexStaf has quit IRC01:34
*** AlexStaf has joined #openstack-lbaas01:50
*** spatel has quit IRC01:54
*** ianychoi has quit IRC02:19
*** ianychoi has joined #openstack-lbaas02:21
*** AlexStaf has quit IRC02:32
*** ramishra has joined #openstack-lbaas02:38
*** ricolin has joined #openstack-lbaas02:40
*** gregwork has joined #openstack-lbaas02:47
*** AlexStaf has joined #openstack-lbaas03:18
*** hongbin has joined #openstack-lbaas03:35
*** psachin has joined #openstack-lbaas03:38
*** hongbin has quit IRC04:22
*** ajay33 has joined #openstack-lbaas04:59
*** gcheresh_ has joined #openstack-lbaas05:00
*** ramishra has quit IRC05:12
*** ramishra has joined #openstack-lbaas05:27
*** gcheresh_ has quit IRC05:33
*** gcheresh_ has joined #openstack-lbaas05:38
*** takamatsu has quit IRC05:54
*** sapd1 has joined #openstack-lbaas06:21
*** pcaruana has joined #openstack-lbaas06:30
*** gcheresh_ has quit IRC06:32
*** vesper11- has quit IRC06:33
*** vesper11 has joined #openstack-lbaas06:36
*** trident has quit IRC06:52
*** trident has joined #openstack-lbaas06:56
*** maciejjozefczyk has joined #openstack-lbaas07:00
*** tesseract has joined #openstack-lbaas07:12
*** rcernin has quit IRC07:16
*** tkajinam has quit IRC07:25
*** tkajinam has joined #openstack-lbaas07:26
*** ataraday has joined #openstack-lbaas07:33
*** ramishra has quit IRC07:47
*** ivve has joined #openstack-lbaas07:51
*** ccamposr has joined #openstack-lbaas07:55
*** tkajinam has quit IRC08:02
*** sapd1 has quit IRC08:04
*** rpittau|afk is now known as rpittau08:07
*** ramishra has joined #openstack-lbaas08:08
*** takamatsu has joined #openstack-lbaas08:26
*** takamatsu has quit IRC08:29
*** takamatsu has joined #openstack-lbaas08:36
*** takamatsu has quit IRC08:41
*** takamatsu has joined #openstack-lbaas08:48
*** lxkong has quit IRC08:49
*** gcheresh_ has joined #openstack-lbaas09:17
*** takamatsu has quit IRC10:56
f0oe:1:/w 311:11
f0osorry11:11
*** takamatsu has joined #openstack-lbaas11:34
*** gcheresh_ has quit IRC11:35
*** takamatsu has quit IRC11:51
*** sapd1 has joined #openstack-lbaas11:51
*** goldyfruit has quit IRC11:58
*** lemko has joined #openstack-lbaas12:00
*** takamatsu has joined #openstack-lbaas12:14
*** takamatsu has quit IRC12:16
*** takamatsu has joined #openstack-lbaas12:21
*** takamatsu has quit IRC12:29
*** takamatsu has joined #openstack-lbaas12:32
*** AustinR has quit IRC12:56
*** takamatsu has quit IRC13:06
*** takamatsu has joined #openstack-lbaas13:10
*** AustinR has joined #openstack-lbaas13:12
*** goldyfruit has joined #openstack-lbaas13:22
*** goldyfruit has quit IRC13:24
*** sapd1 has quit IRC13:28
*** goldyfruit has joined #openstack-lbaas13:32
*** pcaruana has quit IRC13:32
*** takamatsu has quit IRC13:36
openstackgerritAnn Taraday proposed openstack/octavia master: [WIP] Jobboard based controller  https://review.opendev.org/64740613:37
*** KeithMnemonic has joined #openstack-lbaas13:43
*** pcaruana has joined #openstack-lbaas14:04
*** AustinR has quit IRC14:05
*** AustinR has joined #openstack-lbaas14:05
*** ajay33 has quit IRC14:12
*** ivve has quit IRC14:45
*** colin- has joined #openstack-lbaas15:04
*** lemko has quit IRC15:39
*** maciejjozefczyk has quit IRC15:52
johnsomFYI for those involved in the PTG Shanghai: https://etherpad.openstack.org/p/shanghai-ptg-octavia15:58
johnsomThey had a bot create the etherpad links....15:58
johnsomI guess we already had an etherpad. I can update their links page with our page:16:00
johnsomhttps://etherpad.openstack.org/p/octavia-shanghai-U-ptg16:00
johnsomIn case you are looking for the list: http://ptg.openstack.org/etherpads.html16:02
*** rpittau is now known as rpittau|afk16:03
*** henriqueof has joined #openstack-lbaas16:36
*** ivve has joined #openstack-lbaas16:36
*** AustinR has quit IRC16:38
*** AustinR has joined #openstack-lbaas16:44
*** takamatsu has joined #openstack-lbaas16:49
*** AustinR has quit IRC17:08
*** AustinR has joined #openstack-lbaas17:09
*** takamatsu has quit IRC17:12
*** goldyfruit_ has joined #openstack-lbaas17:15
*** goldyfruit has quit IRC17:18
*** ricolin has quit IRC17:20
*** goldyfruit___ has joined #openstack-lbaas17:36
*** psachin has quit IRC17:38
*** goldyfruit_ has quit IRC17:39
*** ramishra has quit IRC17:50
*** goldyfruit_ has joined #openstack-lbaas18:05
*** openstackgerrit has quit IRC18:07
*** goldyfruit___ has quit IRC18:07
*** openstackgerrit has joined #openstack-lbaas18:11
openstackgerritMichael Johnson proposed openstack/octavia master: Add client authentication to the LB cookbook  https://review.opendev.org/68877618:11
openstackgerritMichael Johnson proposed openstack/octavia master: Add client authentication to the LB cookbook  https://review.opendev.org/68877618:12
johnsomI already had a devstack setup for TLS, so figured I could hammer out these cookbook updates quickly....18:13
colin-thanks for the quick fix on the CVE everyone18:19
openstackgerritMichael Johnson proposed openstack/octavia master: Add client authentication to the LB cookbook  https://review.opendev.org/68877618:24
johnsomhelps if I number the outline correctly....18:24
openstackgerritMerged openstack/octavia stable/rocky: loadbalancer vip-network-id IP availability check  https://review.opendev.org/68544718:31
*** gcheresh_ has joined #openstack-lbaas18:47
rm_workjohnsom: did we fix the nova-hv-down issue? assuming running on master, and an amp's HV is down and stays down, can we failover the LB currently?19:06
rm_worki don't remember if we did a fix, or if someone is still working on making it better, or both19:06
johnsomrm_work That is part of the failover flow fix I'm working on19:14
rm_workk19:15
rm_worki thought we might have also put in a hotfix or something that'd workaround the issue somewhat...19:15
johnsomThere is an open nova bug for it (and one for storage)19:15
rm_workany, biab, gotta get dented19:15
rm_work*anyway19:15
*** henriqueof has quit IRC19:21
openstackgerritBrian Haley proposed openstack/octavia master: Stop testing python 2  https://review.opendev.org/68737019:22
*** trident has quit IRC19:29
*** trident has joined #openstack-lbaas19:33
*** gcheresh_ has quit IRC19:35
*** tesseract has quit IRC19:49
*** pcaruana has quit IRC20:25
colin-quick question, in a normal loadbalancer failover flow, why does the BACKUP not takeover MASTER during the period of time it is the only existing amp? (no spares pool, ~1m boot times)20:33
johnsomIt does, but note, the DB is what the SETTING is, not the current state.20:33
colin-ohhh20:34
colin-cisco used to do this with their HA, Active/Standby vs Primary/Secondary20:34
colin-that tripped me up for years20:34
*** pcaruana has joined #openstack-lbaas20:34
colin-got it. thx20:34
johnsomYeah, it seems to catch most folks. We discussed changing the column name, but pretty much nothing would solve it....20:34
colin-in this sense 'role' is actually most appropriate i think. 'status' or something more ephemeral would be what i was thinking of20:35
colin-(who currently has the VIP)20:35
johnsomRight. Plus the amps can failover much faster than we would update the DB.... They are autonomous at that layer of the HA pie20:36
openstackgerritMichael Johnson proposed openstack/octavia master: Add client authentication to the LB cookbook  https://review.opendev.org/68877620:46
*** pcaruana has quit IRC20:51
*** AustinR has quit IRC21:01
rm_workI had a system that caused them to actually update ;D21:07
rm_workso in my deployment at GD, those columns were updated to always be accurate21:07
* rm_work shrugs21:07
rm_workwe could still do that I think21:07
rm_workif we wanted21:08
rm_workit's just informational, so we could do best-effort updates for it21:08
rm_workbut that might be more confusing if it worked "most of the time" than just not doing it at all21:08
rm_workhmm actually i think i know a way to make it always update, theoretically21:08
johnsomRight, I think the confusion and effort would out-weight the benefit.21:08
rm_workI just don't know how it would impact some of our other flow stuff21:08
rm_workthe issue was that keepalived didn't really have a way to query to ask if it was currently acting as master or not21:09
rm_workand also the case of split-brain where both can access the controller21:09
rm_workbecause we COULD track it ourselves using the notify scripts to drop a flag somewhere and then adding that amp status to the UDP health message21:10
johnsomPlus if something is really borked and it's switching like every second, are we just making the cloud bork worse....21:10
rm_workbut I feel like then it'd be much more common to have weird states (hopefully but not guaranteed to be temporarily) where there's two MASTER or BACKUP21:10
rm_workand how that would affect other long-failover stuff...21:11
rm_worki've given it a bit of thought, and there's no COMPLETELY clean way to do it, AFAICT, though there's some options that might be worth the tradeoffs, maybe21:11
johnsomPersonally, I perfectly happy to say "It's autonomous, it changes on it's own in reaction to failures."21:12
johnsomWe could also call it "Quantum HA", it exists in both states, all of the time.21:12
johnsomlol21:13
rm_work>_<21:33
openstackgerritMichael Johnson proposed openstack/octavia master: Add backend re-encryption to the LB cookbook  https://review.opendev.org/68882422:01
johnsomOk, did my tech-debt good deed from the week....22:02
*** goldyfruit_ has quit IRC22:08
*** ivve has quit IRC22:12
*** ccamposr__ has joined #openstack-lbaas22:27
*** ccamposr has quit IRC22:29
*** rcernin has joined #openstack-lbaas22:30
*** tkajinam has joined #openstack-lbaas23:16
johnsomI think retro-fitting my master devstack VM back to queens for testing is a bad idea.... glance seems to 500 out, sigh23:27

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!