| openstackgerrit | Merged openstack/octavia master: Speed up pylint by using multiple cores https://review.openstack.org/637348 | 00:09 |
|---|---|---|
| openstackgerrit | Merged openstack/neutron-lbaas master: Fix the neutron-lbaas-to-octavia-migration job https://review.openstack.org/635056 | 00:12 |
| *** sapd1 has joined #openstack-lbaas | 00:58 | |
| *** Dinesh_Bhor has joined #openstack-lbaas | 01:55 | |
| *** Dinesh_Bhor has quit IRC | 02:11 | |
| *** Dinesh_Bhor has joined #openstack-lbaas | 02:21 | |
| *** psachin has joined #openstack-lbaas | 02:51 | |
| *** sapd1 has quit IRC | 03:07 | |
| *** hongbin has joined #openstack-lbaas | 03:10 | |
| *** sapd1 has joined #openstack-lbaas | 03:23 | |
| *** ramishra has joined #openstack-lbaas | 04:07 | |
| *** sapd1 has quit IRC | 04:48 | |
| *** sapd1 has joined #openstack-lbaas | 05:00 | |
| rm_work | yeah IMO abandon changes from 99cloud and inspur that aren't useful, they are running scripts to try to boost their commit stats, and usually bring no value (and won't respond to anything) | 05:09 |
| *** hongbin has quit IRC | 05:42 | |
| openstackgerrit | Reedip proposed openstack/octavia-tempest-plugin master: Check Monitor in Member Scenario Tests https://review.openstack.org/634891 | 06:06 |
| *** AlexStaf has quit IRC | 06:20 | |
| *** yboaron_ has quit IRC | 07:02 | |
| *** ccamposr has joined #openstack-lbaas | 07:11 | |
| *** ccamposr__ has joined #openstack-lbaas | 07:12 | |
| *** ccamposr has quit IRC | 07:15 | |
| *** gcheresh has joined #openstack-lbaas | 07:20 | |
| *** pcaruana has joined #openstack-lbaas | 07:43 | |
| *** yboaron_ has joined #openstack-lbaas | 08:02 | |
| *** yboaron_ has quit IRC | 08:06 | |
| *** yboaron_ has joined #openstack-lbaas | 08:07 | |
| *** AlexStaf has joined #openstack-lbaas | 08:12 | |
| *** rpittau has joined #openstack-lbaas | 08:17 | |
| cgoncalves | rm_work, your message passed spell check. congrats! | 08:19 |
| rm_work | lol | 08:19 |
| *** fnaval_ has quit IRC | 08:34 | |
| *** fnaval has joined #openstack-lbaas | 08:44 | |
| *** fnaval has quit IRC | 08:49 | |
| *** pcaruana|afk| has joined #openstack-lbaas | 09:01 | |
| *** pcaruana has quit IRC | 09:02 | |
| *** ramishra has quit IRC | 09:05 | |
| *** ramishra has joined #openstack-lbaas | 09:09 | |
| *** ramishra_ has joined #openstack-lbaas | 09:21 | |
| *** ramishra has quit IRC | 09:23 | |
| *** ltomasbo has joined #openstack-lbaas | 09:32 | |
| *** sapd1 has quit IRC | 09:33 | |
| *** trident has joined #openstack-lbaas | 09:38 | |
| openstackgerrit | Reedip proposed openstack/octavia-tempest-plugin master: Check Monitor in Member Scenario Tests https://review.openstack.org/634891 | 09:45 |
| *** sapd1 has joined #openstack-lbaas | 09:49 | |
| *** pcaruana|afk| has quit IRC | 10:01 | |
| *** ramishra_ has quit IRC | 10:06 | |
| *** pcaruana has joined #openstack-lbaas | 10:07 | |
| *** ramishra has joined #openstack-lbaas | 10:09 | |
| openstackgerrit | zhulingjie proposed openstack/neutron-lbaas master: Update hacking version to 1.1.0 https://review.openstack.org/628646 | 10:13 |
| *** ipo has joined #openstack-lbaas | 10:23 | |
| *** numans has joined #openstack-lbaas | 10:24 | |
| ipo | Hello all !!! | 10:24 |
| *** salmankhan has joined #openstack-lbaas | 10:24 | |
| ipo | Please help me to understand more smoothly octavia config parameters for health check | 10:26 |
| ipo | heartbeat_interval is the interval to send HB packets from amphorae. | 10:27 |
| *** salmankhan has quit IRC | 10:29 | |
| ipo | so if octavia didnt received HB packet within heartbeat_timeout from amphora it moves LB to helthy amphora and kill original one ? | 10:33 |
| rm_work | yes | 10:49 |
| rm_work | that's a good summary | 10:49 |
| *** cgoncalves has quit IRC | 10:52 | |
| openstackgerrit | Merged openstack/octavia-tempest-plugin master: Update the live jobs to set higher retries https://review.openstack.org/637074 | 10:55 |
| *** cgoncalves has joined #openstack-lbaas | 10:56 | |
| ipo | ok, in this case what is it health_check_interval ? | 11:14 |
| ipo | is this option connected with amphorae healthcheck ? | 11:15 |
| rm_work | that is deployed TO the amphora's config | 11:19 |
| rm_work | so that's how often it broadcasts its health status packet | 11:19 |
| rm_work | (in seconds) | 11:19 |
| *** Dinesh_Bhor has quit IRC | 11:20 | |
| ipo | I thought it is about heartbeat_interval | 11:20 |
| ipo | it is 10 secs by default | 11:21 |
| rm_work | err hold on | 11:21 |
| rm_work | sorry, half-paying attention and misread that | 11:21 |
| rm_work | yeah | 11:21 |
| rm_work | health_check_interval is how often it actually CHECKS to see if the status is bad for a failover | 11:22 |
| ipo | ok | 11:22 |
| ipo | So for example each 3 secs it check state of amphorae, so after 60 secs of HB timeout in next 3 sec interval healthcheck will find amphora is down ? | 11:22 |
| rm_work | yes | 11:22 |
| ipo | ok, this is clear. Is it the same behavior for SINGLE and ACTIVE_STEND modes ? | 11:25 |
| ipo | so the failover will be managed on the same way (as I remember ACTIVE_STEND will failover much faster), but heartbeat_timeout is 60 secs... | 11:31 |
| *** ipo has quit IRC | 11:35 | |
| *** ipo has joined #openstack-lbaas | 12:03 | |
| *** ipo has quit IRC | 12:09 | |
| openstackgerrit | Nir Magnezi proposed openstack/octavia master: Refactor the unplugging of the VIP https://review.openstack.org/613685 | 13:04 |
| *** trown|outtypewww is now known as trown | 13:06 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia-tempest-plugin master: Add active/standby scenario test https://review.openstack.org/637073 | 13:19 |
| *** yboaron_ has quit IRC | 14:04 | |
| *** yboaron_ has joined #openstack-lbaas | 14:04 | |
| *** psachin has quit IRC | 14:28 | |
| *** fnaval has joined #openstack-lbaas | 15:02 | |
| *** yboaron_ has quit IRC | 15:26 | |
| *** gcheresh has quit IRC | 15:54 | |
| *** AlexStaf has quit IRC | 16:08 | |
| *** pcaruana has quit IRC | 16:10 | |
| *** ramishra has quit IRC | 16:13 | |
| *** ramishra has joined #openstack-lbaas | 16:17 | |
| *** ramishra has quit IRC | 16:42 | |
| cgoncalves | CI is really in a bad mood. grenade queens to rocky runs fine locally, too | 17:00 |
| cgoncalves | http://paste.openstack.org/show/745274/ | 17:00 |
| cgoncalves | that vs http://logs.openstack.org/99/637399/1/check/octavia-grenade/d5d8a62/logs/grenade.sh.txt.gz#_2019-02-16_22_40_11_418 | 17:00 |
| johnsom | That is an interesting one. It implies we have an amp-agent issue, but that is hard to confirm. This is happening in rocky only? | 17:11 |
| johnsom | Are we sure you have the rocky image in your local setup? | 17:11 |
| cgoncalves | stable/rocky only yes, but where it fails it is still pre-upgrade so queens | 17:12 |
| cgoncalves | grenade runs the octavia devstack plugin so it builds the amp (which is queens-based) | 17:12 |
| cgoncalves | running ubuntu xenial, just like in CI | 17:13 |
| johnsom | Oh, ok, so queens image, rocky controller? | 17:13 |
| johnsom | Or queens-queens? | 17:13 |
| cgoncalves | when it fails it still is queens-queens | 17:13 |
| johnsom | hmmmm, ok, now that is odd. | 17:13 |
| cgoncalves | mega odd | 17:14 |
| johnsom | I see our queens gates are passing as of two days ago.... | 17:15 |
| cgoncalves | right, they have always passed | 17:16 |
| cgoncalves | the problem is octavia-grenade in stable/rocky | 17:16 |
| *** rpittau has quit IRC | 17:16 | |
| johnsom | Hmmm, darn, we don't save the DIB logs in that job | 17:16 |
| *** ccamposr__ has quit IRC | 17:18 | |
| *** rpittau has joined #openstack-lbaas | 17:19 | |
| johnsom | cgoncalves Do you have an idea of when this started to fail? | 17:20 |
| cgoncalves | johnsom, december 14-17 | 17:22 |
| johnsom | Hmm, we haven't really touched that area of the agent since August 6th | 17:23 |
| *** rpittau has quit IRC | 17:26 | |
| johnsom | Ah, we did touch os utils the week before that. Let me review that. | 17:27 |
| *** AlexStaf has joined #openstack-lbaas | 17:47 | |
| *** sapd1 has quit IRC | 17:59 | |
| *** trown is now known as trown|lunch | 18:01 | |
| *** trown|lunch is now known as trown | 19:06 | |
| *** hongbin has joined #openstack-lbaas | 19:36 | |
| *** hongbin has quit IRC | 19:43 | |
| *** rcernin has joined #openstack-lbaas | 21:40 | |
| *** trown is now known as trown|outtypewww | 22:02 | |
| *** fnaval has quit IRC | 22:06 | |
| *** fnaval has joined #openstack-lbaas | 22:11 | |
| * johnsom Head bangs... | 22:29 | |
| johnsom | This unset ACLs for barbican is a mess | 22:30 |
| johnsom | Why do we care? | 22:30 |
| johnsom | Right now, if a ref is used as a default cert on one listener and as a sni cert on another, and the default cert listener is deleted, we no longer have access to the cert for SNI use | 22:31 |
| johnsom | Not to mention it's not thread safe in the least | 22:32 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Fix the lose of access to barbican secrets https://review.openstack.org/637646 | 22:51 |
| rm_work | johnsom: erg, forgot to check SNI | 22:57 |
| rm_work | my bad prolly | 22:57 |
| rm_work | checked if it was still in use as a listener cert but not SNI T_T | 22:57 |
| johnsom | Yeah, It's a nasty cross use problem that is going to get worse with the more TLS features. I think it is best to just leave the ACL in place | 22:58 |
| rm_work | BTW, is it possible to control how many threads are used for unit test running? | 22:58 |
| rm_work | like in tox.ini maybe | 22:58 |
| rm_work | johnsom: erg, that's... i mean i guess... but... eugh | 22:58 |
| johnsom | I think so | 22:58 |
| rm_work | not exactly the best security feeling | 22:58 |
| johnsom | Well, as it is written right now it's not thread safe either | 22:59 |
| rm_work | i guess it's running stestr | 22:59 |
| rm_work | so i'd need to add stestr args | 22:59 |
| rm_work | ah yeah true :/ | 23:00 |
| johnsom | --concurrency=2 | 23:00 |
| johnsom | on stestr | 23:00 |
| johnsom | You need to dial it back? | 23:01 |
| rm_work | yeah | 23:01 |
| rm_work | this is for nova, i'm hitting test issues in our gate which runs like freaking 16 at once | 23:01 |
| rm_work | where one thread mysteriously dies every time | 23:01 |
| johnsom | Yeah, so with the bug, the threading issue, and that it has no test coverage, my vote is to nuke it | 23:02 |
| rm_work | trying to see if dialing it back to what my machine does locally helps | 23:02 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Fix the loss of access to barbican secrets https://review.openstack.org/637646 | 23:03 |
| rm_work | yeah i ... can't really object | 23:04 |
| rm_work | i'm not sure how we'd fix a concurrency issue | 23:04 |
| rm_work | can't lock something we don't even know about | 23:04 |
| johnsom | Yeah, that is what makes it a bit hard | 23:04 |
| rm_work | since it could be adding the same cert to a different LB | 23:05 |
| rm_work | and we can't know that | 23:05 |
| rm_work | i guess we could have like a.... | 23:05 |
| rm_work | cert lock table | 23:05 |
| rm_work | ugh that gets messy lol | 23:05 |
| johnsom | Right, you could create a critical region in the code, but.... | 23:05 |
| rm_work | yeah it's not the best, but i think i'm with you] | 23:05 |
| rm_work | though i believe someone could say it's laziness at the cost of security <_<] | 23:06 |
| rm_work | technically | 23:06 |
| johnsom | I would not argue that with them | 23:08 |
| johnsom | Though I think the risk is fairly low | 23:08 |
| rm_work | yes | 23:10 |
| rm_work | which is why i'm not up in arms | 23:10 |
| rm_work | there's ... worse problems | 23:11 |
| * rm_work shrugs | 23:11 | |
| rm_work | though that's not ideal logic <_< | 23:11 |
| rm_work | broken window? i forget | 23:11 |
| rm_work | ah no that's a financial thing | 23:12 |
| johnsom | broken arrow? | 23:12 |
| johnsom | grin | 23:12 |
| rm_work | lol | 23:12 |
| johnsom | I should have the bulk of the TLS client auth code cleaned up today. | 23:15 |
| rm_work | k, you'll be wanting reviews i guess :P | 23:17 |
| rm_work | poke me at the start of the chain when you're ready | 23:17 |
| johnsom | I want a bunch of reviews. I'm going to re-organize the review list today/tomorrow by deadlines | 23:18 |
| johnsom | Next week is non-client libraries, week after is feature freeze for everything else | 23:18 |
| rm_work | kk | 23:19 |
| rm_work | wow already? | 23:19 |
| johnsom | Yep | 23:19 |
| rm_work | <_< | 23:19 |
| johnsom | If you have a few minutes: https://review.openstack.org/#/q/project:openstack/octavia-lib | 23:20 |
| eandersson | johnsom, rm_work lets get this merged :p https://review.openstack.org/#/c/636428/ | 23:21 |
| eandersson | So I can rebase tomorrow when we are back in the office =] | 23:21 |
| rm_work | i'm not going to vote on it :P other than +1 | 23:22 |
| rm_work | i think your nit is correct tho <_< | 23:22 |
| rm_work | tempted to fix it... | 23:23 |
| eandersson | haha | 23:23 |
| eandersson | do it | 23:23 |
| rm_work | i mean, should i also have "if Transport is not None" for the others? | 23:25 |
| rm_work | i guess that is ... | 23:25 |
| rm_work | fine | 23:25 |
| eandersson | Yea | 23:25 |
| rm_work | well | 23:25 |
| rm_work | hmmm | 23:25 |
| rm_work | yea | 23:25 |
| eandersson | but not as important | 23:25 |
| rm_work | running tests locall | 23:27 |
| rm_work | y | 23:27 |
| rm_work | kk | 23:28 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix oslo messaging connection leakage https://review.openstack.org/636428 | 23:28 |
| johnsom | easndersson FYI, the oslo folks are also looking at if they can make a change on the messaging side to improve the situation. | 23:30 |
| johnsom | It was discussed at today's meeting | 23:30 |
| eandersson | Yea - they highlighted me a few times | 23:30 |
| rm_work | fixed | 23:30 |
| rm_work | so i'll leave it up to another core to +2 | 23:30 |
| rm_work | ah and cgoncalves needs to refresh his | 23:30 |
| johnsom | I am a bit concerned about what else isn't getting closed in those API threads. | 23:31 |
| rm_work | hmm | 23:33 |
| rm_work | eandersson: the thing that tipped it for me BTW was this: https://docs.openstack.org/bandit/latest/plugins/b101_assert_used.html | 23:34 |
| rm_work | i mentioned it earlier in chat but forgot to post on the patch | 23:35 |
| rm_work | i should do that for posterity | 23:35 |
| johnsom | Yeah, that would be good | 23:35 |
| rm_work | done | 23:36 |
| rm_work | which ... thanks VSCode for showing me that error there, pycharm hadn't been | 23:36 |
| rm_work | but vscode includes all the bandit results inline :P | 23:37 |
| rm_work | which is pretty cool | 23:37 |
| johnsom | Joy, somehow I have managed to make flake8 crash.... | 23:38 |
| eandersson | Sure - but assertion can be fixed like this https://github.com/openstack/neutron-lib/blob/master/neutron_lib/rpc.py#L74 | 23:38 |
| rm_work | johnsom: i think even more important than pylint concurrency might be flake8 concurrency <_< | 23:40 |
| rm_work | might solve it being stupidly slow for me, testing | 23:40 |
| eandersson | It's not really arguing the change that you emphasized on with your "singleton" pattern | 23:40 |
| rm_work | eandersson: asserts there may not work | 23:41 |
| eandersson | but honestly I am so over this dicussion | 23:41 |
| eandersson | You guys do what ever you want | 23:41 |
| rm_work | they are completely ignored when compiled to bytecode | 23:41 |
| *** eandersson has left #openstack-lbaas | 23:41 | |
| rm_work | ah yeah, that is a workaround | 23:45 |
| johnsom | Hmm, flake8 doesn't seem too bad to me. | 23:46 |
| johnsom | Their -j flag doesn't seem to do much for me. The pylint helps for me. | 23:47 |
| johnsom | But I think they are all mostly IO bound to some degree | 23:47 |
| rm_work | hmm i'm on SSD, you? | 23:48 |
| johnsom | Some SSD some NAS | 23:48 |
| rm_work | hmm, the flake8 step doesn't print time | 23:49 |
| rm_work | so i'm gonna run timing tests on my own now | 23:49 |
| rm_work | yeah even with -j0 my CPU doesn't really increase with pep8 tests | 23:50 |
| rm_work | :/ | 23:50 |
| johnsom | On pylint with -j 0 it lights up my cores | 23:51 |
| rm_work | hmm | 23:53 |
| rm_work | yeah i feel like it isn't working | 23:54 |
| rm_work | i still see only one python process using one core | 23:54 |
| johnsom | -j on flake8 seems to do nothing | 23:54 |
| rm_work | right | 23:54 |
| johnsom | flake8 for me gives: real 0m25.128s | 23:54 |
| rm_work | lol | 23:55 |
| rm_work | real 0m45.509s | 23:56 |
| johnsom | pylint with the -j change: real 0m35.466s | 23:56 |
| rm_work | ah --benchmark | 23:56 |
| rm_work | is nice | 23:56 |
| rm_work | yeah -j does not seem to be doing anything | 23:57 |
| johnsom | Without -j on pylint: real 1m10.667s | 23:57 |
| rm_work | ah it defaults to auto anyway | 23:58 |
| rm_work | weird | 23:58 |
| johnsom | Yeah, but if I even put in -j 6 on flake, it doesn't use them | 23:59 |
| rm_work | right | 23:59 |
| rm_work | noticing the same | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!