Tuesday, 2019-01-29

rm_workwell, I +2'd the first 400:05
rm_workreading through the doc one00:05
rm_workthey don't look egregiously bad so they'll do :P00:05
johnsomI am basically at the point of stopping development due to the backlog00:05
rm_workthat's why we need to get this stuff through and then tweak00:05
rm_worklike i said last week, we have time00:05
rm_workto fix any bugs that fall out from testing00:05
rm_workthe overhead is getting pretty insane00:06
rm_workand anyway, we need more low-hanging fruit for newbies to dive into, so we can get some new folks in here :)00:06
johnsomFunny enough I have been cleaning up some tech debt as I went along....00:06
rm_workthere's more though :P00:08
rm_workbut i stopped thinking THAT deep into it after the first one00:08
rm_workonce i have infinite time, i'll go back and refactor the things i noted there00:08
rm_worki'm so close to cracking that time-loop thing00:09
rm_workjohnsom: i can't believe there's already bugfix patches in this chain... wat00:12
rm_workyou didn't want to just ... fix it?00:12
johnsomWell, they were edge cases and the chain was already a bit long-ish00:12
johnsomI have been trying to make the individual patches not too large....00:13
johnsomDidn't think you all wanted another one of those 5,000 line patches....00:13
rm_worki JUST thought about spares+flavors00:13
rm_worksomehow it escaped me00:14
rm_workuntil now00:14
rm_workanywho, everything up to there is +2'd00:14
johnsomWell, me too..... Yang called me on it00:14
rm_worklooking for the tempest tests for this chain00:15
johnsomDid you miss this one? https://review.openstack.org/#/c/63259400:16
johnsomTempest is here00:16
rm_worki said i +2'd everything "up to" that one00:16
rm_worki need to think00:16
rm_workgod the tempest test comments are so verbose00:17
rm_worki get why00:17
rm_workbecause they get used in IDEs00:17
rm_workbut like00:17
rm_workdo you hate my "Super-DRY" _create_object / _update_object / etc abstractions yet? :P00:18
johnsomI'm a bit frustrated with the number of "clients" and SDKs we have in OpenStack00:18
johnsomNo, they are awesome. Though I had to hack one cleanup as something didn't have the right status fields.  Amps I think.00:19
rm_work+2'd the tempest side client additions at least, gonna read the test side in a sec00:19
rm_workah yeah there should have been an example for that00:19
rm_worki dunno, every time i look back at shit like that i did, i wonder if it's more weird for people to use than it adds in DRYness00:20
johnsomJust overrode it, no big deal00:20
rm_workI think "yeah, that was clever", but "wait, is that a bad thing"00:20
rm_workbrb afk a bit <_<00:20
*** hongbin has joined #openstack-lbaas00:27
*** trown is now known as trown|outtypewww00:39
rm_workso you need another +2 :P01:07
rm_work... several01:08
*** yamamoto has joined #openstack-lbaas01:09
*** hongbin has quit IRC01:18
*** hongbin has joined #openstack-lbaas01:20
*** dulek has quit IRC01:28
*** Dinesh_Bhor has joined #openstack-lbaas01:32
*** yamamoto has quit IRC02:03
*** yamamoto has joined #openstack-lbaas02:41
*** yamamoto has quit IRC02:46
*** psachin has joined #openstack-lbaas02:52
*** aojea has joined #openstack-lbaas03:12
*** aojea has quit IRC03:17
*** yamamoto has joined #openstack-lbaas03:25
*** yamamoto has quit IRC03:34
openstackgerritMerged openstack/octavia-tempest-plugin master: Add the flavor profile service client.  https://review.openstack.org/63040503:35
*** Dinesh_Bhor has quit IRC03:43
*** Dinesh_Bhor has joined #openstack-lbaas03:50
*** yamamoto has joined #openstack-lbaas04:06
*** hongbin has quit IRC06:21
*** gcheresh has joined #openstack-lbaas06:23
*** ramishra has joined #openstack-lbaas06:33
*** strigazi has quit IRC06:42
*** strigazi has joined #openstack-lbaas06:43
*** ccamposr has joined #openstack-lbaas06:53
*** aojea has joined #openstack-lbaas06:57
*** aojea has quit IRC07:07
openstackgerritYang JianFeng proposed openstack/octavia master: Some adjustments about l3-active-active blueprint  https://review.openstack.org/63194707:27
*** yamamoto has quit IRC07:28
*** yamamoto has joined #openstack-lbaas07:32
openstackgerritYang JianFeng proposed openstack/octavia master: Some adjustments about l3-active-active blueprint  https://review.openstack.org/63194707:34
*** sapd1 has joined #openstack-lbaas07:48
*** rpittau has joined #openstack-lbaas08:04
*** dulek has joined #openstack-lbaas08:15
*** ramishra_ has joined #openstack-lbaas08:34
*** ramishra has quit IRC08:35
*** yamamoto has quit IRC08:35
*** yamamoto has joined #openstack-lbaas08:36
openstackgerritYang JianFeng proposed openstack/octavia master: Some adjustments about l3-active-active blueprint  https://review.openstack.org/63194708:40
*** pcaruana has joined #openstack-lbaas08:51
jitekaHello, I've got a question about LBs topology09:44
jitekaCould it be possible to migrate active STANDALONE LBs to ACTIVE_STANDBY with a failover ?09:44
jitekaI wonder if update queries into octavia DB amphora table to change STANDALONE amphora with role MASTER would do the trick09:44
jitekaThanks in advance :)09:44
*** numans_ is now known as numans_afk09:56
*** numans_afk is now known as numans10:07
openstackgerritMerged openstack/octavia master: Add flavor, flavor_profile table and their APIs  https://review.openstack.org/48649910:13
openstackgerritMerged openstack/octavia master: Add flavors/flavor_profile api-ref  https://review.openstack.org/61832210:13
openstackgerritMerged openstack/octavia master: Adds flavor support to the amphora driver  https://review.openstack.org/62132310:13
openstackgerritMerged openstack/octavia master: Add provider driver capabilities API  https://review.openstack.org/62419010:13
*** yboaron has joined #openstack-lbaas10:20
*** ramishra_ has quit IRC10:23
*** sapd1 has quit IRC10:23
*** ramishra has joined #openstack-lbaas10:30
*** Dinesh_Bhor has quit IRC10:38
*** mkuf_ has quit IRC11:07
*** mkuf has joined #openstack-lbaas11:07
*** yamamoto has quit IRC11:14
*** sapd1 has joined #openstack-lbaas11:17
*** sapd1 has quit IRC11:36
*** yamamoto has joined #openstack-lbaas11:54
*** yamamoto has quit IRC12:04
*** mkuf has quit IRC12:20
*** yamamoto has joined #openstack-lbaas12:23
*** mkuf has joined #openstack-lbaas12:26
*** pcaruana has quit IRC12:40
*** yboaron has quit IRC12:42
*** pcaruana has joined #openstack-lbaas12:50
*** kaiokmo has joined #openstack-lbaas12:51
*** mkuf_ has joined #openstack-lbaas12:59
*** mkuf has quit IRC13:02
*** aojea_ has joined #openstack-lbaas13:10
*** trown|outtypewww is now known as trown13:13
*** aojea_ has quit IRC13:14
*** abaindur has joined #openstack-lbaas13:39
*** gcheresh_ has joined #openstack-lbaas13:43
*** mkuf has joined #openstack-lbaas13:43
*** gcheresh has quit IRC13:44
*** mkuf_ has quit IRC13:45
*** abaindur has quit IRC13:51
*** psachin has quit IRC13:56
*** yamamoto has quit IRC14:00
jitekaAlso another question, just found out in operator documentation that it's now possible to perform amphora failover14:05
jitekaalso found corresponding python-octaviaclient release note include that command14:05
jitekaAny ETA about when or if it will be backported to Rocky version ?14:05
jitekaI found relative change in gerrit I'm talking about : https://review.openstack.org/#/c/532424/14:08
*** aojea_ has joined #openstack-lbaas14:12
*** yboaron has joined #openstack-lbaas14:14
*** yamamoto has joined #openstack-lbaas14:23
openstackgerritMerged openstack/python-octaviaclient master: Add --provider to load balancer create  https://review.openstack.org/63356214:29
*** aojea_ has quit IRC14:30
*** aojea_ has joined #openstack-lbaas14:33
*** pcaruana has quit IRC14:45
*** pcaruana has joined #openstack-lbaas14:53
johnsomjiteka: yes, in theory you can hack the DB and failover to an active/standby LB. Though I have not tested it.14:55
johnsomjiteka: per the OpenStack stable policy, we cannot backport features to stable releases.14:56
*** salmankhan has joined #openstack-lbaas15:14
*** yamamoto has quit IRC15:15
*** aojea_ has quit IRC15:16
*** yamamoto has joined #openstack-lbaas15:16
*** yamamoto has quit IRC15:16
*** aojea_ has joined #openstack-lbaas15:16
*** yamamoto has joined #openstack-lbaas15:17
*** salmankhan has quit IRC15:18
*** aojea_ has quit IRC15:21
*** yamamoto has quit IRC15:23
*** gcheresh_ has quit IRC15:39
*** gcheresh_ has joined #openstack-lbaas15:40
*** gcheresh_ has quit IRC15:45
*** openstackgerrit has quit IRC15:51
*** gcheresh_ has joined #openstack-lbaas16:07
*** gcheresh_ has quit IRC16:15
xgermanjiteka: In the Stein version we will have flavors which will make it easier to migrate Standalone->Active/Standby16:51
*** ccamposr has quit IRC17:01
*** pcaruana has quit IRC17:01
*** yamamoto has joined #openstack-lbaas17:21
*** rpittau has quit IRC17:21
*** rpittau has joined #openstack-lbaas17:21
colin-lxkong: i happened to find your name in an ansible commit related to the openstack cloud module it provides17:23
colin-(a comment in os_loadbalancer is what i'm referring to) and i wondered if you know whether os_quota for type loadbalancer refers to the octavia resource or if it's the old neutron one?17:23
*** yboaron has quit IRC17:24
*** yboaron has joined #openstack-lbaas17:24
*** ramishra has quit IRC17:25
*** yboaron has quit IRC17:26
*** trown is now known as trown|lunch17:26
*** yamamoto has quit IRC17:27
*** yboaron has joined #openstack-lbaas17:27
*** yboaron has quit IRC17:31
*** yboaron has joined #openstack-lbaas17:31
*** yboaron has quit IRC17:35
*** rpittau has quit IRC17:35
jitekajohnsom: Will give it a try it then to see how it behave but I will keep in mind what german said about flavor with stein18:00
jitekajohnsom xgerman : thanks you both18:00
johnsomLooks like the merge chain is pending the client patch merge:  https://review.openstack.org/#/c/624294/18:06
johnsomDue to docs that reference the client changes.18:06
*** aojea has joined #openstack-lbaas18:10
*** aojea has quit IRC18:15
*** trown|lunch is now known as trown18:21
rm_workjiteka: i have done that topo change, and it is somewhat as you describe, though extra steps are involved18:28
*** aojea has joined #openstack-lbaas18:29
rm_workjohnsom: +A'd18:30
johnsomThank you18:37
*** gcheresh_ has joined #openstack-lbaas18:46
*** aojea has quit IRC18:52
*** aojea has joined #openstack-lbaas18:53
cgoncalvesis LBs in DELETED status more of an amphora specific thing or you see the value/need of having such status for other providers?18:56
*** aojea has quit IRC18:57
cgoncalvesasking because if also applicable to other providers, non-amphora deployments also kind of depend on o-hk to clean up deleted LBs18:57
johnsomIt is important that provider drivers send the "DELETED" status back to the controller, not only for tracking, but it will also trigger the VIP port delete if the Octavia API created the port for the driver..18:58
cgoncalveso-hk is somewhat coupled with the amphora driver, where it executes other tasks like spare pool check and amphora expire18:59
johnsomYep, we talked about that. Either deploying the HM, ignoring them, or doing something custom.18:59
cgoncalvess/HM/HK, right?18:59
johnsomRight, sorry, HK...18:59
rm_workyeah i thought HK was part of a full deployment regardless of provider19:00
rm_workit isn't amphora specific19:00
johnsomYeah, if you deploy with no Octavia, you might want to purge those, might not.19:00
rm_workjust happens to have a few amphora tasks in it19:00
cgoncalvesfor the most part it is today19:00
cgoncalveswould be nice to have a way in HK to tell if the amphora driver is enabled. if not, don't run amphora specific tasks19:01
johnsomBasically, three out of the four functions HK has would not be needed with only third party deployments.19:01
cgoncalvesbut that is not easy to tell19:01
johnsomBut, it's one line of SQL and cron to replace that anyway, so.....19:01
rm_workright since providers mean multiple drivers are enabled...19:01
cgoncalvesbecause either HK would need to consume same config file as API or have duplicated enabled_provider_drivers19:02
rm_workit also isn't a huge amount of overhead...19:02
rm_workbut not ideal19:02
johnsomIs there really enough load to worry about with HK running and only doing DB purge?  It's not like it runs all the time or does much, especially if you don't have amp records19:02
cgoncalvestotally agree19:03
cgoncalvesjust wanted to be sure that my understanding was correct and not missing something important19:03
rm_workwat: http://logs.openstack.org/94/624294/9/gate/test-release-openstack/d804b41/job-output.txt.gz19:04
*** yamamoto has joined #openstack-lbaas19:04
cgoncalvesas long as they stay as threads.......19:04
johnsomIt's going to wake up, query the DB, then go back to sleep....19:04
cgoncalvesrm_work, right? :P19:04
rm_worki think so yeah19:04
rm_workjust not a huge deal19:04
rm_workso wtf happened to https://review.openstack.org/#/c/624294/19:04
cgoncalvesrm_work, I was teasing you over the switch from threads to processes in HM by you19:05
rm_workah lol19:05
rm_worktoo many threads (of conversation)19:05
johnsomDid he add an import and not include it in requirements?  How did the gates pass.... hmmmm19:06
johnsomNope, I don't see where he added OpenSSL....  Plus, I don't think he needs it19:06
cgoncalvesdoesn't seem related to the patch itself19:06
johnsomAh, twine....19:07
cgoncalvesspeaking of not including requirements https://review.openstack.org/#/c/633359/19:09
johnsomThat is the "new" way to publish to pypi19:10
johnsomNot an openstack tool19:10
rm_workso err19:14
rm_worktwine is broken and all of openstack publishing will fail?19:14
* rm_work guesses19:15
*** fnaval has joined #openstack-lbaas19:39
*** gcheresh_ has quit IRC19:53
*** yamamoto has quit IRC20:05
*** blake has joined #openstack-lbaas20:05
johnsomUgh, so the recheck failed....  I'm going to look into that gate....20:14
*** openstackgerrit has joined #openstack-lbaas20:21
openstackgerritAdam Harwell proposed openstack/python-octaviaclient master: Add python3-openssl to bindep for twine  https://review.openstack.org/63382020:21
*** blake has quit IRC20:25
*** aojea has joined #openstack-lbaas20:29
*** aojea has quit IRC20:30
*** aojea_ has joined #openstack-lbaas20:30
openstackgerritMichael Johnson proposed openstack/python-octaviaclient master: Add provider command support  https://review.openstack.org/62429421:20
*** fnaval has quit IRC21:28
*** trown is now known as trown|outtypewww22:01
*** aojea_ has quit IRC22:33
-openstackstatus- NOTICE: http://zuul.openstack.org is not working. https://zuul.openstack.org does work. Please use that while we investigate.23:13
openstackgerritMerged openstack/python-octaviaclient master: Add provider command support  https://review.openstack.org/62429423:35
johnsomCool time to +W bump the chain....23:35
openstackgerritMerged openstack/octavia master: Add flavor docs and releae notes  https://review.openstack.org/62681923:52
*** rcernin has quit IRC23:53

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