| *** armax has joined #openstack-lbaas | 00:30 | |
| *** armax has quit IRC | 00:38 | |
| *** reedip has quit IRC | 00:56 | |
| *** armax has joined #openstack-lbaas | 01:02 | |
| *** armax has quit IRC | 01:22 | |
| *** SumitNaiksatam has left #openstack-lbaas | 01:41 | |
| *** gongysh has joined #openstack-lbaas | 01:49 | |
| *** reedip has joined #openstack-lbaas | 01:53 | |
| *** blogan has quit IRC | 02:11 | |
| *** aleph1 has quit IRC | 02:38 | |
| *** reedip has quit IRC | 02:43 | |
| *** aleph1 has joined #openstack-lbaas | 02:47 | |
| *** raginbajin has quit IRC | 02:47 | |
| *** raginbajin has joined #openstack-lbaas | 02:52 | |
| *** yuanying has quit IRC | 02:58 | |
| *** reedip has joined #openstack-lbaas | 03:09 | |
| *** fnaval has quit IRC | 03:13 | |
| *** fnaval has joined #openstack-lbaas | 03:14 | |
| *** gongysh has quit IRC | 03:20 | |
| *** reedip has quit IRC | 03:43 | |
| *** reedip has joined #openstack-lbaas | 03:48 | |
| *** fnaval has quit IRC | 04:00 | |
| *** links has joined #openstack-lbaas | 04:20 | |
| *** yuanying has joined #openstack-lbaas | 04:20 | |
| *** amotoki has joined #openstack-lbaas | 04:31 | |
| *** amotoki has quit IRC | 04:35 | |
| *** gcheresh_ has joined #openstack-lbaas | 04:40 | |
| *** amotoki has joined #openstack-lbaas | 04:44 | |
| *** reedip has quit IRC | 04:50 | |
| *** reedip has joined #openstack-lbaas | 04:54 | |
| *** amotoki has quit IRC | 04:58 | |
| *** amotoki has joined #openstack-lbaas | 05:01 | |
| *** gcheresh_ has quit IRC | 05:04 | |
| *** amotoki has quit IRC | 05:09 | |
| *** armax has joined #openstack-lbaas | 05:11 | |
| *** gongysh has joined #openstack-lbaas | 05:29 | |
| *** eandersson_ has joined #openstack-lbaas | 05:38 | |
| *** reedip has quit IRC | 05:38 | |
| *** reedip has joined #openstack-lbaas | 05:42 | |
| *** eandersson has quit IRC | 05:42 | |
| *** amotoki has joined #openstack-lbaas | 05:58 | |
| *** reedip has quit IRC | 06:05 | |
| *** cpuga has joined #openstack-lbaas | 06:06 | |
| *** aojea has joined #openstack-lbaas | 06:12 | |
| *** aojea has quit IRC | 06:18 | |
| *** armax has quit IRC | 06:29 | |
| *** reedip has joined #openstack-lbaas | 06:42 | |
| *** gongysh has quit IRC | 06:51 | |
| *** gongysh has joined #openstack-lbaas | 07:02 | |
| *** eezhova has joined #openstack-lbaas | 07:29 | |
| *** tesseract has joined #openstack-lbaas | 07:31 | |
| openstackgerrit | zoushilin proposed openstack/octavia master: Remove unused logging import https://review.openstack.org/449455 | 07:38 |
|---|---|---|
| *** cpuga has quit IRC | 07:53 | |
| *** gcheresh_ has joined #openstack-lbaas | 07:54 | |
| *** gcheresh_ has quit IRC | 08:00 | |
| *** gcheresh_ has joined #openstack-lbaas | 08:08 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/neutron-lbaas-dashboard master: Imported Translations from Zanata https://review.openstack.org/449474 | 08:09 |
| *** eezhova has quit IRC | 08:17 | |
| *** aojea has joined #openstack-lbaas | 08:18 | |
| *** aojea has quit IRC | 08:18 | |
| *** aojea has joined #openstack-lbaas | 08:19 | |
| *** gongysh has quit IRC | 08:22 | |
| *** gongysh has joined #openstack-lbaas | 08:38 | |
| rm_work | Ah, I have some quick-notes for zero-to-octavia install | 08:39 |
| rm_work | if that is useful | 08:39 |
| rm_work | since i've been working on a deployment and i restart from scratch A LOT | 08:39 |
| rm_work | and need it to be consistent | 08:40 |
| *** gongysh has quit IRC | 08:43 | |
| *** reedip has quit IRC | 08:49 | |
| *** reedip has joined #openstack-lbaas | 08:50 | |
| *** eezhova has joined #openstack-lbaas | 08:51 | |
| *** gcheresh_ has quit IRC | 08:55 | |
| openstackgerrit | zoushilin proposed openstack/octavia master: Remove unused logging import https://review.openstack.org/449455 | 09:02 |
| *** reedip has quit IRC | 09:03 | |
| *** reedip has joined #openstack-lbaas | 09:04 | |
| *** aojea_ has joined #openstack-lbaas | 09:10 | |
| *** yuanying has quit IRC | 09:12 | |
| *** aojea has quit IRC | 09:12 | |
| *** gongysh has joined #openstack-lbaas | 09:20 | |
| *** reedip has quit IRC | 09:41 | |
| openstackgerrit | huangshan proposed openstack/neutron-lbaas master: Remove log translations from neutron-lbaas https://review.openstack.org/447921 | 09:48 |
| *** yamamoto has quit IRC | 09:50 | |
| *** yuanying has joined #openstack-lbaas | 09:54 | |
| *** pcaruana has joined #openstack-lbaas | 10:05 | |
| *** gcheresh_ has joined #openstack-lbaas | 10:25 | |
| *** reedip has joined #openstack-lbaas | 10:28 | |
| openstackgerrit | Bernard Cafarelli proposed openstack/octavia master: DIB: drop some dependencies https://review.openstack.org/448657 | 10:34 |
| *** gongysh has quit IRC | 10:37 | |
| *** yamamoto has joined #openstack-lbaas | 10:49 | |
| *** yamamoto_ has joined #openstack-lbaas | 10:50 | |
| *** yamamoto has quit IRC | 10:54 | |
| *** armax has joined #openstack-lbaas | 11:02 | |
| *** yamamoto_ has quit IRC | 11:05 | |
| *** yamamoto has joined #openstack-lbaas | 11:05 | |
| *** armax has quit IRC | 11:16 | |
| *** yamamoto has quit IRC | 11:17 | |
| openstackgerrit | He Qing proposed openstack/neutron-lbaas master: Fix exception about loadbalancer_stats https://review.openstack.org/449573 | 11:19 |
| *** gongysh has joined #openstack-lbaas | 11:28 | |
| *** fyxim has joined #openstack-lbaas | 12:00 | |
| *** links has quit IRC | 12:02 | |
| *** links has joined #openstack-lbaas | 12:05 | |
| *** yamamoto has joined #openstack-lbaas | 12:06 | |
| *** sputnik13 has joined #openstack-lbaas | 12:10 | |
| *** yamamoto has quit IRC | 12:15 | |
| *** gongysh has quit IRC | 12:17 | |
| reedip | o/ | 12:20 |
| *** yamamoto has joined #openstack-lbaas | 13:19 | |
| xgerman | o/ | 13:34 |
| *** links has quit IRC | 13:53 | |
| *** cpuga has joined #openstack-lbaas | 14:07 | |
| rm_work | o/ | 14:23 |
| rm_work | WTB merging things today plz | 14:24 |
| rm_work | i might look at getting pagination/sorting on top of the new API since now i can do the root tag stuff | 14:24 |
| *** chlong has joined #openstack-lbaas | 14:25 | |
| *** cpuga_ has joined #openstack-lbaas | 14:27 | |
| *** cpuga has quit IRC | 14:31 | |
| *** amotoki has quit IRC | 14:35 | |
| *** m-greene_ has quit IRC | 14:44 | |
| *** amotoki has joined #openstack-lbaas | 14:45 | |
| *** m-greene has quit IRC | 14:46 | |
| rm_work | diltram: wait, the lxd tests pass now?! | 14:46 |
| *** amotoki has quit IRC | 14:47 | |
| *** m-greene has joined #openstack-lbaas | 14:48 | |
| *** m-greene_ has joined #openstack-lbaas | 14:48 | |
| openstackgerrit | huangshan proposed openstack/octavia master: Remove log translations from octavia https://review.openstack.org/447872 | 14:49 |
| rm_work | augh new DIB issue | 14:50 |
| rm_work | 2017-03-23 21:28:55.799 | ImportError: No module named diskimage_builder.paths | 14:50 |
| rm_work | with py3 only | 14:50 |
| rm_work | though it works locally... | 14:54 |
| rm_work | something is wonky | 14:54 |
| diltram | rm_work: no, it's not LXD real job | 15:01 |
| diltram | it's just normal scenario job currently | 15:02 |
| diltram | I'm working on it | 15:02 |
| diltram | but we're finally gonna have working this gate job as they upstream change into cloud-init | 15:03 |
| diltram | and now I need just to test it | 15:03 |
| diltram | and mark as working and they will upstream it | 15:03 |
| rm_work | ahhh i see | 15:04 |
| rm_work | also ahhh i see to the dib issue | 15:04 |
| rm_work | their script has no care for python_version | 15:04 |
| rm_work | T_T | 15:04 |
| rm_work | need to fix this again T_T | 15:04 |
| rm_work | greghaynes: https://github.com/openstack/diskimage-builder/blob/master/diskimage_builder/lib/disk-image-create#L34 | 15:06 |
| rm_work | greghaynes: this doesn't respect devstack's python version at all T_T | 15:07 |
| rm_work | and I feel like there are probably many more places with the same problem | 15:07 |
| *** blogan has joined #openstack-lbaas | 15:10 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Install DIB from source so depends-on will work https://review.openstack.org/444707 | 15:16 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Install DIB from source so depends-on will work https://review.openstack.org/444707 | 15:16 |
| rm_work | I feel like this is going to fail, then I can submit the fix and depends-on it | 15:16 |
| rm_work | but want to wait for the demonstrable fail first | 15:17 |
| johnsom | rm_work on the listener patch, you said you were removing the "protocol/protocol_port from the listener PUT right? | 15:21 |
| rm_work | didn't i already? | 15:21 |
| johnsom | Did you change get accidentally reverted? | 15:22 |
| rm_work | errrg | 15:22 |
| rm_work | yeah i think maybe | 15:22 |
| rm_work | i did a lot of work here | 15:22 |
| rm_work | that looks to be gone | 15:22 |
| rm_work | ugh | 15:22 |
| johnsom | sad face | 15:22 |
| rm_work | ok sec | 15:22 |
| rm_work | i can fix | 15:22 |
| rm_work | hmmm | 15:23 |
| rm_work | I wonder if i didn't even push that one up? | 15:23 |
| rm_work | ugh sec | 15:23 |
| rm_work | gotta find it | 15:23 |
| rm_work | i know i did this | 15:23 |
| rm_work | oh | 15:24 |
| rm_work | eugh wtf | 15:25 |
| rm_work | this is going to drive me crazy | 15:28 |
| rm_work | johnsom: ok so... keeping it will add validation burden | 15:33 |
| rm_work | otherwise i'm not against leaving it there? | 15:33 |
| rm_work | but | 15:33 |
| rm_work | i don't know if anyone would ever RELY on not being able to change it? >_> | 15:33 |
| johnsom | I don't think we should allow it personally. | 15:33 |
| rm_work | k | 15:37 |
| rm_work | i'll throw up another patchset | 15:37 |
| johnsom | Is anyone else seeing this? https://bugs.launchpad.net/octavia/+bug/1675795 | 15:38 |
| openstack | Launchpad bug 1675795 in octavia "Octavia diskimage-create utility failing due to latest changes in diskimage-builder repo" [Undecided,New] | 15:38 |
| johnsom | My last devstacks ran fine. I wondering if this is against a stable branch | 15:39 |
| rm_work | err | 15:40 |
| rm_work | so in py35 it fails | 15:40 |
| rm_work | and i am LITERALLY RIGHT NOW pushing up a fix | 15:40 |
| rm_work | let me see if it's the same thing | 15:40 |
| rm_work | ah not the same thing | 15:40 |
| johnsom | Does it fail that way? | 15:40 |
| johnsom | Ok | 15:40 |
| johnsom | Ok, I asked for more information. It is grey and rainy here, I'm going to go make a coffee run. Back in a few. | 15:42 |
| rm_work | k | 15:42 |
| *** fnaval has joined #openstack-lbaas | 15:52 | |
| *** amotoki has joined #openstack-lbaas | 15:56 | |
| *** pcaruana has quit IRC | 16:03 | |
| rm_work | johnsom: errr do you remember what the assertEqual for lists that DOESN'T care about order is? | 16:27 |
| rm_work | i guess i can just sort them | 16:27 |
| johnsom | Is it assertIn? | 16:28 |
| johnsom | I guess that doesn't mean equivalence though | 16:29 |
| johnsom | assertItemsEqual but I think that is py3 | 16:30 |
| rm_work | ah yeah | 16:30 |
| johnsom | Ah, it is in 2.7 as well | 16:31 |
| rm_work | I think it might be backported to py27 | 16:31 |
| rm_work | yeah | 16:31 |
| *** eezhova has quit IRC | 16:31 | |
| rm_work | in fact I worked on that | 16:31 |
| rm_work | also did some test work | 16:33 |
| rm_work | tls tests work now | 16:34 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Octavia v2 API for listeners https://review.openstack.org/424744 | 16:34 |
| rm_work | tls was a little broken | 16:35 |
| rm_work | should work now | 16:35 |
| johnsom | Cool. I'm looking at if DIB broke our stable branches again. I think so. | 16:36 |
| rm_work | lol | 16:36 |
| rm_work | hmm wtf tho | 16:36 |
| rm_work | why is MY patch passing py3x scenario gates on dib-master | 16:36 |
| rm_work | but your patch isn't passing them on pypi-dib | 16:36 |
| rm_work | which includes the same patches | 16:36 |
| rm_work | <_< | 16:36 |
| rm_work | 2.1.0 | 16:37 |
| rm_work | ugh i'll worry about py3x-dib later | 16:37 |
| rm_work | http://logs.openstack.org/57/438757/9/check/gate-octavia-v1-dsvm-py3x-scenario-ubuntu-xenial-nv/15b749e/logs/devstacklog.txt.gz#_2017-03-23_21_28_55_799 | 16:38 |
| rm_work | I believe this is the fix: https://review.openstack.org/#/c/449721/ | 16:39 |
| rm_work | but I am not sure why it's *passing* against master (which is basically 2.1.0) but not against pypi (2.1.0) | 16:40 |
| *** cpuga_ has quit IRC | 16:41 | |
| rm_work | johnsom: https://github.com/openstack/diskimage-builder/commit/02e5353a9fd5bba77162d6eff43c22917773da3d | 16:42 |
| rm_work | maybe that will fix the issue when released to pypi? | 16:42 |
| rm_work | it JUST didn't make the cut for 2.1.0 i think | 16:42 |
| rm_work | oh nm it did | 16:43 |
| rm_work | hmm then... whelp | 16:43 |
| johnsom | I think I need to work on something else as my frustration level is getting a bit high | 16:44 |
| rm_work | something else, as in.... | 16:44 |
| rm_work | go spin some clay and make some pottery? | 16:45 |
| rm_work | or just not look at DIB :P | 16:45 |
| diltram | johnsom: ping | 16:46 |
| *** cpuga has joined #openstack-lbaas | 16:47 | |
| johnsom | Not look at DIB issues, like go test listener API stuff | 16:47 |
| johnsom | diltram pong | 16:47 |
| diltram | I have a question about our devstack/plugin.sh | 16:49 |
| diltram | what was idea behind using this vars names | 16:50 |
| diltram | $OCTAVIA_DIR/diskimage-create/diskimage-create.sh $octavia_dib_tracing_arg -o $OCTAVIA_AMP_IMAGE_FILE ${PARAM_OCTAVIA_AMP_BASE_OS:-''} ${PARAM_OCTAVIA_AMP_DISTRIBUTION_RELEASE_ID:-''} ${PARAM_OCTAVIA_AMP_IMAGE_SIZE:-''} | 16:50 |
| diltram | with PARAM_ prefix? | 16:50 |
| diltram | do you know something about? | 16:50 |
| diltram | or I can change it | 16:50 |
| diltram | because I'm working on this rhel | 16:50 |
| diltram | and I want to export some vars to make it working | 16:51 |
| johnsom | It was this recent patch: https://review.openstack.org/#/c/431576/ | 16:51 |
| diltram | I know | 16:51 |
| diltram | but why the prefix | 16:51 |
| diltram | this PARAM_ | 16:51 |
| diltram | may I drop this prefix | 16:52 |
| johnsom | I think the point was for OCTAVIA_AMP_BASE_OS to be a flag and PARAM_OCTAVIA_AMP_BASE_OS to contain the command line strings "-i centos" | 16:52 |
| johnsom | I.e. gate hook would set OCTAVIA_AMP_BASE_OS =True | 16:52 |
| rm_work | johnsom: errr, so are *all* PUT actions in v2 returning status 200? | 16:53 |
| rm_work | or just some? | 16:53 |
| rm_work | does LB do status 202? | 16:53 |
| johnsom | Wait, 200??? | 16:53 |
| rm_work | apparently? | 16:54 |
| johnsom | diltram if you overwrite OCTAVIA_AMP_BASE_OS with the string "-i centos" it would work, but you couldn't do boolean tests on it anywhere else in the plugin.sh | 16:54 |
| rm_work | i dunno someone changed this recently | 16:54 |
| johnsom | rm_work let me check the reference... Just a minute | 16:55 |
| diltram | ok but I should set OCTAVIA_AMP_BASE_OS=centos | 16:55 |
| diltram | and in settings I can set default values | 16:56 |
| diltram | like ubuntu | 16:56 |
| *** aojea_ has quit IRC | 16:56 | |
| diltram | and always run script with param | 16:56 |
| rm_work | johnsom: was changed by ankur-gupta-f1 in patch 31 | 16:56 |
| rm_work | ankur-gupta-f1: did you mention somewhere why PUT status changed from 202 to 200? | 16:56 |
| rm_work | ankur-gupta-f4: ^^ in case | 16:57 |
| johnsom | Oh, yes. The defaults should work ok for the others. If you want to be specific, put in 7 for the centos release to make sure it uses 7 and not 6. | 16:57 |
| ankur-gupta-f4 | neutron-lbaas | 16:57 |
| rm_work | ankur-gupta-f4: ok so you are sure n-lbaas does status 200 for PUT? | 16:57 |
| ankur-gupta-f4 | No wait | 16:57 |
| ankur-gupta-f4 | That sounds wrong | 16:58 |
| ankur-gupta-f4 | Ugh | 16:58 |
| ankur-gupta-f4 | Double checking | 16:58 |
| rm_work | yes it SOUNDS wrong, but | 16:58 |
| rm_work | neutron-lbaas is a dumpster fire | 16:58 |
| rm_work | so i don't doubt it | 16:58 |
| johnsom | rm_work, yep, nlbaas does 200 for PUTs | 16:59 |
| johnsom | At least on LB | 16:59 |
| rm_work | yeah | 16:59 |
| rm_work | FML | 16:59 |
| ankur-gupta-f4 | Oh good | 16:59 |
| rm_work | I hate everything and everyone | 16:59 |
| rm_work | anywho... | 17:00 |
| johnsom | rm_work what would you use? | 17:00 |
| rm_work | 202 | 17:00 |
| rm_work | because it's async | 17:00 |
| johnsom | Yeah. Most of these are DB changes, so instant, but yeah | 17:00 |
| rm_work | uhh | 17:01 |
| rm_work | ok i guess SOME are | 17:01 |
| rm_work | but I feel like the majority are not? | 17:01 |
| rm_work | dunno, i'd have to count | 17:01 |
| johnsom | We don't let folks "update" much | 17:01 |
| rm_work | hmm | 17:01 |
| johnsom | Yeah, no worries. | 17:01 |
| johnsom | Well, in queens you can add v3 that just fixes all of the status codes.... grin | 17:02 |
| openstackgerrit | Lubosz Kosnik (diltram) proposed openstack/octavia master: Update DIB config in plugin.sh https://review.openstack.org/449733 | 17:03 |
| *** cpuga has quit IRC | 17:05 | |
| diltram | johnsom, rm_work: ^^ | 17:05 |
| johnsom | diltram I now remember why we didn't do what you just posted. It means we have default values in two places. | 17:05 |
| diltram | I know but we should keep defaults in settings | 17:06 |
| diltram | this is why this file exist :P | 17:06 |
| diltram | so I can make this differently | 17:06 |
| johnsom | No | 17:06 |
| johnsom | Those settings are for devstack, not for our image builder | 17:06 |
| diltram | I know | 17:06 |
| rm_work | diltram: umm, if we're not using PARAM_ anymore, shouldn't you completely remove them? | 17:06 |
| diltram | I can make this in normal way | 17:07 |
| diltram | that if something is set I can append this to command | 17:07 |
| diltram | rm_work: where it's left | 17:07 |
| rm_work | diltram: posted | 17:09 |
| rm_work | johnsom: and i agree with diltram here :/ | 17:09 |
| johnsom | rm_work I don't like having defaults stored in two places. When we update from xenial we will have to go change both or the gates using devstack would run a different image than someone just running the script | 17:10 |
| rm_work | well it depends | 17:10 |
| diltram | johnsom: exporting vars with anything more than name (xenial,trusty,7,6) is awful | 17:11 |
| rm_work | will our devstack defaults always equal what we think the user will deploy with by default? | 17:11 |
| rm_work | I don't think so | 17:11 |
| rm_work | I think there could be variance | 17:11 |
| openstackgerrit | Merged openstack/octavia master: Use HostAddressOpt for opts that accept IP and hostnames https://review.openstack.org/447989 | 17:11 |
| diltram | rm_work: agree | 17:11 |
| diltram | ok nvm | 17:12 |
| diltram | I didn't saw this previous lines | 17:12 |
| diltram | abandoned patch | 17:14 |
| *** catintheroof has joined #openstack-lbaas | 17:14 | |
| *** catintheroof has quit IRC | 17:15 | |
| *** catintheroof has joined #openstack-lbaas | 17:15 | |
| rm_work | ah? | 17:18 |
| rm_work | aight | 17:18 |
| *** yamamoto has quit IRC | 17:19 | |
| diltram | johnsom: https://review.openstack.org/449741 ^^ | 17:21 |
| johnsom | diltram We can't mix "RHEL" and "centos".... | 17:23 |
| diltram | ok | 17:23 |
| johnsom | Otherwise looks fine to me | 17:24 |
| diltram | done | 17:26 |
| johnsom | I updated the topic and +1'd | 17:27 |
| diltram | thx | 17:27 |
| *** eezhova has joined #openstack-lbaas | 17:29 | |
| johnsom | diltram tox didn't like your project-config patch | 17:32 |
| diltram | I see :( | 17:32 |
| diltram | johnsom: do you know how to use those vars in configuration file? | 17:35 |
| johnsom | Like which? | 17:36 |
| diltram | I thought that I'm gonna prepare it for future but don't know how to refer those values | 17:36 |
| diltram | like this {hypervisor} and based on this I would like to enable LXD or not | 17:36 |
| johnsom | OCTAVIA_AMP_BASE_OS=centos | 17:36 |
| johnsom | OCTAVIA_AMP_DISTRIBUTION_RELEASE_ID=7 | 17:36 |
| johnsom | Those? | 17:36 |
| diltram | also | 17:36 |
| diltram | like based on centos7 in gate check name I would like to add those | 17:37 |
| johnsom | Oh, | 17:37 |
| ankur-gupta-f4 | johnsom: anything to target/do today specifically. | 17:41 |
| johnsom | Get listener merged I think | 17:41 |
| rm_work | johnsom: uhh, if we delete a pool and the delete fails | 17:46 |
| rm_work | should the listener be in ERROR status or ONLINE? | 17:46 |
| johnsom | pool should be in provisioning ERROR, listener would be provis ACTIVE, listener operating status should turn.... ERROR I think | 17:47 |
| johnsom | I guess it depends on where in the delete it failed. | 17:48 |
| rm_work | handler | 17:49 |
| johnsom | Operating status is the observed status. So if it failed before it actually deleted, it would still show online as technically it is still online and handling traffic | 17:49 |
| *** tesseract has quit IRC | 17:50 | |
| rm_work | hmmmm | 17:51 |
| rm_work | for some reason these tests are passing when run alone and failing when run as a group >_> | 17:51 |
| ankur-gupta-f4 | How does one run the scenario tests in octavia? | 17:53 |
| johnsom | ankur-gupta-f4 Have a devstack up and running with tempest, export TEMPEST_CONFIG_DIR=/opt/stack/tempest/etc, then tox -e scenario | 17:54 |
| *** yamamoto has joined #openstack-lbaas | 17:56 | |
| *** reedip has quit IRC | 17:56 | |
| openstackgerrit | Ankur proposed openstack/octavia master: Introduce TESTING.rst for quick developer reference https://review.openstack.org/403938 | 17:57 |
| rm_work | almost done with a pools update | 17:59 |
| *** cpuga has joined #openstack-lbaas | 18:11 | |
| *** cpuga has quit IRC | 18:17 | |
| *** cpuga_ has joined #openstack-lbaas | 18:18 | |
| *** jiahao has joined #openstack-lbaas | 18:18 | |
| *** yamamoto has quit IRC | 18:19 | |
| jiahao | Hi team, quick question. If I am going to install octavia from stable/newton, which version of neutron-lbaas-dashboard should I use? It seems there is no stable/newton for neutron-lbaas-dashboard. | 18:20 |
| johnsom | jiahao You can use stable/mitaka for newton Octavia | 18:21 |
| rm_work | johnsom: is it too late to just CUT ONE for newton? | 18:22 |
| rm_work | we get this question about once a week | 18:22 |
| johnsom | No, we could still do that.... It's just odd | 18:23 |
| johnsom | I do regret not just cutting one... That was my mistake.... | 18:24 |
| *** chlong has quit IRC | 18:24 | |
| jiahao | But based on the release cycle, anything on stable/mitaka is unmaintained, correct? | 18:25 |
| johnsom | Well, technically it still gets security updates for a few more weeks. | 18:26 |
| *** eezhova has quit IRC | 18:27 | |
| johnsom | It's just that there were not real changes in the newton cycle to the dashboard, so there wasn't much point to cut one. You can also just install the 1.0.0 version from pypi | 18:27 |
| jiahao | ok. I see. Thanks for clarifying | 18:29 |
| rm_work | johnsom: erm, do pools only have one status? not operating / provisioning? | 18:32 |
| rm_work | seems wrong | 18:32 |
| johnsom | No, every object has both | 18:32 |
| johnsom | Should.... | 18:32 |
| rm_work | k | 18:32 |
| johnsom | In Octavia, not nlbaas | 18:32 |
| rm_work | oh | 18:33 |
| rm_work | in n-lbaas it only has one? | 18:33 |
| rm_work | that's what i meant | 18:33 |
| johnsom | Well, nlbaas doesn't return those on each API call, it's only the status tree | 18:34 |
| rm_work | erm | 18:34 |
| rm_work | we are now | 18:34 |
| rm_work | in lb and listeners | 18:34 |
| rm_work | is that wrong? | 18:34 |
| johnsom | No, it's the way forward. It's kind of dumb to not return that | 18:35 |
| rm_work | ok... | 18:35 |
| rm_work | so pools SHOULD return that | 18:35 |
| johnsom | Ok, so I just checked, with status tree pool has both | 18:35 |
| rm_work | in v2 | 18:35 |
| rm_work | ok but, just the pool return | 18:36 |
| johnsom | on n-lbaas | 18:36 |
| rm_work | not the status tree | 18:36 |
| johnsom | Right, just pool calls, I don't think it returns both, maybe not either of them | 18:36 |
| rm_work | ok, but LB and listeners DO? | 18:37 |
| rm_work | on just the LB / listener return, not status tree | 18:37 |
| rm_work | fff, mock trouble | 18:45 |
| johnsom | Who is working on the listener patch? | 18:45 |
| johnsom | The tenant_id isn't returning | 18:46 |
| sshank | johnsom, I think thats cause the return from POST is return directly from self._send_listener_to_handler()? | 18:48 |
| johnsom | rm_work Sorry, had a call. Did you want me to test something on neutron-lbaas? | 18:48 |
| rm_work | johnsom: GET for LB / listener / pool | 18:49 |
| johnsom | sshank Yeah, I am wondering if we didn't get the "copy project_id" code applied to the listener patch | 18:49 |
| rm_work | and tell me if it's really just "pool" that returns an object with "status" instead of "operating_status" and "provisioning_status" | 18:49 |
| rm_work | johnsom: i'll look in a sec | 18:49 |
| rm_work | i've almost got the pools patch done | 18:50 |
| rm_work | sorting out a mock issue | 18:50 |
| johnsom | https://www.irccloud.com/pastebin/EVIZpd50/ | 18:50 |
| johnsom | neutron-lbaas get LB | 18:51 |
| rm_work | ok | 18:51 |
| johnsom | https://www.irccloud.com/pastebin/ipowCIuo/ | 18:51 |
| rm_work | so it does | 18:51 |
| johnsom | nlbaas listener show | 18:52 |
| rm_work | NO status | 18:52 |
| rm_work | interesting | 18:52 |
| rm_work | none at all | 18:52 |
| rm_work | so ... we CAN still return it though if we want? :P | 18:52 |
| rm_work | (i think we want) | 18:52 |
| johnsom | https://www.irccloud.com/pastebin/8DqH0FvV/ | 18:53 |
| johnsom | nlbaas pool show | 18:53 |
| johnsom | So that is all neutron-lbaas v2 API. In Octavia v2 API it will be on EVERY object | 18:53 |
| ankur-gupta-f4 | rm_work: johnsom Who is taking care of the tenant_id issue in Listeners | 18:54 |
| johnsom | Well, health monitor doesn't have operating status for example. | 18:54 |
| johnsom | ankur-gupta-f4 I asked that question a few minutes ago... Grin | 18:54 |
| ankur-gupta-f4 | yea i didnt see an answer | 18:55 |
| johnsom | I kind of didn't want to stop testing to go fix that | 18:55 |
| sshank | I think sindhu was doing listeners? | 18:55 |
| ankur-gupta-f4 | rm_work: are you going to take care of it or are you too deep into the other one? | 18:56 |
| ankur-gupta-f4 | sshank: we have gotten to a stage where we are just trying to take care of things as soon as they come up instead of waiting for the original author to address comments. | 18:57 |
| sshank | Ok. | 18:57 |
| johnsom | Other than that, create looks good | 18:57 |
| johnsom | I need to come back and test barbican. I forgot to add it to my localrc for this stack... | 18:58 |
| ankur-gupta-f4 | sshank: you want to take care of the tenant_id issue on listeners | 18:58 |
| ankur-gupta-f4 | problem: on POST to listeners, if 'tenant_id:####' it provided, it should also be returned in the Response | 18:59 |
| johnsom | I don't think so, I think the issue is it should be returned on all responses | 18:59 |
| sshank | Ok.I'll give it a try now. | 19:00 |
| johnsom | sshank My guess is this line is missing: https://github.com/openstack/octavia/blob/master/octavia/api/v2/types/load_balancer.py#L56 | 19:01 |
| ankur-gupta-f4 | indeed that would be it | 19:03 |
| *** cpuga_ has quit IRC | 19:03 | |
| ankur-gupta-f4 | sshank: once you add that line also please adjust the listeners controller get_all method | 19:04 |
| sshank | Ok | 19:05 |
| ankur-gupta-f4 | https://github.com/openstack/octavia/blob/master/octavia/api/v2/controllers/load_balancer.py#L57 | 19:05 |
| ankur-gupta-f4 | Should be more like that to accept project/tenant id and then determine based on passed in values or the context | 19:05 |
| ankur-gupta-f4 | thanks sshank | 19:06 |
| rm_work | yeah k i'll pull listeners down again after you fix it then | 19:09 |
| rm_work | i have other stuff to tweak but | 19:09 |
| rm_work | i am deep into pools... | 19:09 |
| rm_work | this mock thing is killing me | 19:09 |
| rm_work | when you run the tests individually, it works | 19:09 |
| rm_work | but running ALL tests, it doesn't | 19:09 |
| rm_work | the mock is different | 19:09 |
| *** kobis1 has joined #openstack-lbaas | 19:10 | |
| *** kobis has quit IRC | 19:11 | |
| johnsom | ankur-gupta-f4 We take the project_id and tenant_id on create listener, but do nothing with it. The listener inherits the project from the load balancer. Right? | 19:13 |
| rm_work | i thought the _type_to_model_map thing took care of it | 19:13 |
| rm_work | in fact what i was looking into momentarily is related | 19:14 |
| johnsom | rm_work This is the patch for v1 that fixed all of that: https://review.openstack.org/#/c/431207/ | 19:15 |
| johnsom | Sindhu fixed the listeners patch for this change already | 19:15 |
| sshank | johnsom, I was think the problem was here: https://review.openstack.org/#/c/424744/32/octavia/api/v2/controllers/listener.py@187 We should pass it inside listener_types.ListenerRootResponse as done in PUT | 19:16 |
| rm_work | yeah | 19:16 |
| johnsom | From what I see, it is just the response is missing the tenant_id | 19:16 |
| rm_work | sshank: the return value of that function *is* a ListenerRootResponse | 19:16 |
| rm_work | so it should be the same | 19:17 |
| sshank | Hmm. Ok. | 19:17 |
| *** yamamoto has joined #openstack-lbaas | 19:17 | |
| rm_work | johnsom: so is the idea that pools don't have a provisioning status? O_o just listeners? | 19:23 |
| rm_work | seems maybe that does make a LITTLE sense | 19:23 |
| rm_work | as a pool is a logical lump on a listener | 19:23 |
| rm_work | a pool on a LB isn't provisioned until it's also on a listener | 19:23 |
| rm_work | and the status is dependent on the specific listener | 19:24 |
| johnsom | rm_work, going forward, everything has provisioning status | 19:24 |
| rm_work | hmmm | 19:24 |
| johnsom | I'm worried we are confusing this.... The pastes I sent were legacy neutron | 19:25 |
| rm_work | i'm actually not sure what that means for pools | 19:25 |
| rm_work | how does a pool have a provisioning status | 19:25 |
| *** cpuga has joined #openstack-lbaas | 19:25 | |
| rm_work | it could be provisioned to any number of different things | 19:25 |
| johnsom | provisioning status, for pools, means did the CRUD operation work. I.e. was the AMP updated or not. | 19:25 |
| rm_work | or not at all | 19:25 |
| rm_work | right? | 19:25 |
| rm_work | hmm | 19:25 |
| rm_work | ok... | 19:25 |
| *** yamamoto has quit IRC | 19:25 | |
| rm_work | so it DOES need one | 19:25 |
| rm_work | fff I'm going to do that later | 19:25 |
| rm_work | only so much concentration right now | 19:26 |
| sshank | When I add those tenant_id stuff, I get weird import base errors. | 19:26 |
| rm_work | also i'm about to not care about this mock thing | 19:26 |
| rm_work | and just skip these tests | 19:26 |
| johnsom | The point is, if an update, for example, of a pool fails, the pool will go into provisioning status ERROR which indicates to the user, something when wrong. | 19:26 |
| johnsom | Then with it in ERROR they have the option to delete and try again to recover. | 19:26 |
| johnsom | Or of course, call up fanatical support to fix whatever went wrong.... Grin | 19:27 |
| rm_work | sshank: let me look at it really quick... pushing up my pool work now | 19:27 |
| sshank | rm_work, Ok | 19:27 |
| rm_work | i'm preparing a hat for consumption presently | 19:28 |
| rm_work | but | 19:28 |
| rm_work | if we don't merge at least listeners today, I'll eat my hat | 19:28 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for pools https://review.openstack.org/405922 | 19:31 |
| rm_work | sshank: I think it's just the one line... | 19:33 |
| rm_work | seems to work for me | 19:33 |
| sshank | Nothing to do in get_all? | 19:34 |
| sshank | get_all of listeners. | 19:34 |
| johnsom | We have the True vs true issue for admin_state_up as well | 19:35 |
| johnsom | Not sure if load balancer is still broken for that or if it was fixed. There is a bug for LB. | 19:35 |
| rm_work | hmm | 19:36 |
| johnsom | nlbaas takes either, this patch is rejecting True | 19:36 |
| rm_work | ugh k | 19:36 |
| rm_work | can we fix that later at the same time as we fix it in LB? | 19:37 |
| johnsom | Sure, I can open a bug for it | 19:37 |
| sshank | rm_work, Yeah. Its only in get_all. | 19:37 |
| johnsom | Updated the bug for True | 19:38 |
| rm_work | ah k | 19:38 |
| rm_work | i thought you said there was one already for LB | 19:38 |
| johnsom | Yeah, I just added listener to the LB bug | 19:39 |
| johnsom | I.e. fix them all | 19:39 |
| rm_work | ummm | 19:45 |
| rm_work | ah nm | 19:45 |
| rm_work | ok i see | 19:45 |
| ankur-gupta-f4 | the get_all is a different issue and needs to be done too | 19:48 |
| ankur-gupta-f4 | rm_work: | 19:48 |
| ankur-gupta-f4 | sshank: | 19:48 |
| rm_work | yeah | 19:48 |
| rm_work | i got it | 19:48 |
| ankur-gupta-f4 | k sorry left comp for a sec and came back and was totally lost | 19:48 |
| ankur-gupta-f4 | any idea where the args get parsed when a POST is sent? | 19:49 |
| ankur-gupta-f4 | i.e. when I try the "True" it returns the error. Where is that error being generated from/by | 19:49 |
| johnsom | Ok, other than those, I think we are ok. I'm going to grab lunch. My last concern is are we doing get all correctly based on admin/non-admin and filters (which probably aren't implemeneted yet) | 19:49 |
| ankur-gupta-f4 | rm_work: Do you know the answer to the above question ^^^ | 19:54 |
| rm_work | lol the test for get_all() for LB was actually broken | 20:00 |
| rm_work | funny | 20:00 |
| rm_work | anyway, will have a patch to push in a sec, johnsom / sshank | 20:00 |
| sshank | rm_work, Ok | 20:00 |
| *** eezhova has joined #openstack-lbaas | 20:00 | |
| *** cpuga has quit IRC | 20:05 | |
| rm_work | yeah actually get_all() was broken in load_balancers >_> | 20:07 |
| rm_work | fixed it | 20:07 |
| rm_work | johnsom: just double-checking, the functionality of get_all() is: | 20:13 |
| rm_work | if admin: either get *everything* if no project_id is provided, or else get based on the project_id provided | 20:13 |
| rm_work | if not admin: only get things for the token's project_id | 20:14 |
| rm_work | is that correct? | 20:14 |
| *** aojea has joined #openstack-lbaas | 20:18 | |
| rm_work | johnsom: | 20:21 |
| ankur-gupta-f4 | correct | 20:21 |
| ankur-gupta-f4 | what is it doing now? | 20:21 |
| johnsom | rm_work: trying to get a bite. Admin should get all, they can use to filters to narrow. Non-admin is constrained to project_id and again can filter more if needed | 20:23 |
| rm_work | right | 20:23 |
| rm_work | kk | 20:23 |
| rm_work | wait | 20:23 |
| rm_work | how would they filter mode | 20:23 |
| rm_work | *more | 20:23 |
| johnsom | Name= | 20:23 |
| *** yamamoto has joined #openstack-lbaas | 20:23 | |
| rm_work | oh, the filtering patch, right | 20:23 |
| rm_work | not included yet | 20:23 |
| johnsom | Right | 20:23 |
| johnsom | Correct | 20:24 |
| rm_work | ok | 20:24 |
| *** armax has joined #openstack-lbaas | 20:29 | |
| *** yamamoto has quit IRC | 20:33 | |
| johnsom | Ok, back from lunch | 20:36 |
| johnsom | I'm going to read through once more. Once the tenant_id is fixed, and if it all looks good I'm ready to +2 | 20:39 |
| ankur-gupta-f4 | sshank: are you going to push up soon? | 20:39 |
| rm_work | hmmmm | 20:43 |
| rm_work | so this is really interesting | 20:43 |
| rm_work | trying to figure out how to test auth/noauth in functional tests | 20:44 |
| rm_work | because i need to prove that it works when you come in as a user who isn't admin | 20:44 |
| rm_work | versus one who is | 20:44 |
| rm_work | do we have a way to do that? | 20:44 |
| johnsom | Ah, it also has the null vs "" issue for name/description | 20:44 |
| rm_work | need to somehow give pecan a context object | 20:44 |
| rm_work | johnsom: listeners does? | 20:45 |
| rm_work | k | 20:45 |
| ankur-gupta-f4 | johnsom: i fixed in my patch | 20:45 |
| ankur-gupta-f4 | Or it should have | 20:45 |
| johnsom | ankur-gupta-f4 Have you pushed or ??? | 20:45 |
| johnsom | I'm running 36bc31b6380557259e2afeda4402efdb4de4c393 | 20:45 |
| rm_work | ankur-gupta-f4: you fixed for LB? | 20:45 |
| rm_work | not Listeners? | 20:45 |
| ankur-gupta-f4 | i fixed both | 20:46 |
| ankur-gupta-f4 | johnsom: patchset 31 | 20:46 |
| johnsom | Yeah, I think LB got fixed, but listener not | 20:46 |
| rm_work | oh, no, i see it | 20:46 |
| rm_work | it's there | 20:46 |
| rm_work | oh | 20:46 |
| rm_work | wrong spot | 20:46 |
| rm_work | i'll fix | 20:46 |
| johnsom | {"listeners": [{"insert_headers": {}, "project_id": "335eb128f3c74848a5632996eb81e268", "protocol": "HTTP", "description": null, "admin_state_up": false, "loadbalancers": [{"id": "96b0fc54-9811-4fa3-b94c-497c60abcc92"}], "created_at": "2017-03-24T20:42:57", "provisioning_status": "ACTIVE", "connection_limit": 200, "updated_at": "2017-03-24T20:42:58", | 20:46 |
| johnsom | "default_pool_id": null, "sni_container_refs": [], "protocol_port": 80, "id": "e64f75b8-fe9b-417c-9727-9e45eda93bfe", "operating_status": "OFFLINE", "name": null}]} | 20:46 |
| *** gcheresh_ has quit IRC | 20:46 | |
| rm_work | yeah | 20:47 |
| rm_work | i see the issue johnsom, fixed it | 20:47 |
| rm_work | it's too far down | 20:47 |
| rm_work | johnsom: so this get_all issue is complicated | 20:47 |
| rm_work | i've got it "fixed" | 20:48 |
| rm_work | but testing is complex | 20:48 |
| johnsom | Oh, this looks ugly too | 20:48 |
| johnsom | {"debuginfo": "Traceback (most recent call last):\n\n File \"/usr/local/lib/python2.7/dist-packages/wsmeext/pecan.py\", line 85, in callfunction\n result = f(self, *args, **kwargs)\n\n File \"/opt/stack/octavia/octavia/api/v2/controllers/listener.py\", line 187, in post\n return self._send_listener_to_handler(context.session, db_listener)\n\n File | 20:48 |
| johnsom | \"/opt/stack/octavia/octavia/api/v2/controllers/listener.py\", line 159, in _send_listener_to_handler\n session, db_listener.id,\n\nAttributeError: 'NoneType' object has no attribute 'id'\n", "faultcode": "Server", "faultstring": "'NoneType' object has no attribute 'id'"} | 20:48 |
| rm_work | what is that from | 20:48 |
| johnsom | Duplicate listener | 20:48 |
| rm_work | k | 20:48 |
| openstackgerrit | German Eichberger proposed openstack/octavia master: [WIP] Adds the ability to failover a load balancer https://review.openstack.org/444347 | 20:49 |
| xgerman | now I need to figure out if it actually works… | 20:50 |
| *** reedip has joined #openstack-lbaas | 20:51 | |
| reedip | ... | 20:51 |
| rm_work | johnsom: soooooo | 20:53 |
| rm_work | johnsom: any idea how to get a project_id into the *pecan context* in the functional tests | 20:53 |
| rm_work | i'm a little stuck | 20:53 |
| rm_work | this should work | 20:53 |
| rm_work | but | 20:53 |
| rm_work | i don't know how to prove it with a test | 20:53 |
| johnsom | Like API functional? | 20:54 |
| rm_work | yes | 20:54 |
| rm_work | ah maybe got it | 20:56 |
| johnsom | rm_work you can't mock octavia/common/context? | 20:57 |
| rm_work | err | 20:57 |
| rm_work | mayhaps :P | 20:57 |
| rm_work | yep | 20:57 |
| rm_work | easier | 20:58 |
| rm_work | hmm i don't think it actually worked tho | 20:59 |
| rm_work | prolly a location issue | 20:59 |
| rm_work | need to figure out WHERE to mock it | 20:59 |
| *** links has joined #openstack-lbaas | 21:05 | |
| *** aojea has quit IRC | 21:06 | |
| johnsom | Ok, bugs in for stats/status/stats | 21:09 |
| ankur-gupta-f4 | k | 21:09 |
| ankur-gupta-f4 | i thought I fixed listeners for null name/description. Exact same method. And I thought when i tested it was fine. But i guess not. Now im trying to figure out why it doesn't hit my pdb in the types file D: | 21:10 |
| rm_work | eugh i lose the session tho | 21:11 |
| rm_work | ankur-gupta-f4: it's too low | 21:11 |
| rm_work | ankur-gupta-f4: just needs to move up above the "if not children" | 21:11 |
| rm_work | ankur-gupta-f4: i fixed it locally, almost ready to push up a bunch of fixes | 21:11 |
| rm_work | if i can get this test to work | 21:12 |
| rm_work | ffff | 21:12 |
| rm_work | so if I mock out the context, i lose the session | 21:12 |
| rm_work | which means nothing works | 21:12 |
| ankur-gupta-f4 | ah k coolio | 21:15 |
| johnsom | Coolio? Are you also doing the Friday 80's hits radio channel??? Grin | 21:16 |
| rm_work | GOT IT k | 21:16 |
| johnsom | Oh, memory fail, coolio was a thing in the 90's.... | 21:16 |
| rm_work | with mock.patch.object(octavia.common.context.Context, 'project_id', listener3['project_id']) as mock_context: | 21:17 |
| rm_work | funtimes | 21:17 |
| johnsom | Yeah, those are always a pain to get right/working | 21:18 |
| *** ndahiwade has quit IRC | 21:21 | |
| rm_work | OK | 21:21 |
| rm_work | i think this is it | 21:21 |
| rm_work | but | 21:21 |
| rm_work | this is a pretty big change maybe T_T | 21:21 |
| *** aojea has joined #openstack-lbaas | 21:22 | |
| rm_work | running tests | 21:22 |
| rm_work | let's see if i forgot anything | 21:22 |
| rm_work | rewrote most of the testing around get_all in both listener and loadbalancer :/ | 21:23 |
| rm_work | it's possible the LB fixes belong in another patchset | 21:23 |
| rm_work | but | 21:23 |
| rm_work | uhh | 21:23 |
| rm_work | i am not sure I care | 21:23 |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas master: Fix a releasenotes issue with rst markup. https://review.openstack.org/449821 | 21:24 |
| rm_work | actually nm it's easy | 21:24 |
| johnsom | ^^^ that is a super easy review to fix a reno issue | 21:24 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Octavia v2 API for listeners https://review.openstack.org/424744 | 21:25 |
| rm_work | johnsom: ok... that is not so simple review ^^ | 21:26 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix get_all method for v2 LB controller https://review.openstack.org/449822 | 21:26 |
| rm_work | that one is more clear ^^ as to what i had to change for get_all | 21:26 |
| rm_work | if you could make sure my thought process there is correct | 21:26 |
| rm_work | that'd be super | 21:27 |
| ankur-gupta-f4 | headed out for the day. Will catch up with IRC later. Still stuck on the true/True issue since I dont know where to start. Good work today yalls | 21:28 |
| rm_work | yeah maybe I can do True/true this weekend | 21:28 |
| rm_work | we'll see my motivation level <_< | 21:28 |
| rm_work | have a good weekend ankur-gupta-f4! :) | 21:28 |
| johnsom | rm_work so start with LB get_all? | 21:29 |
| rm_work | yeah | 21:29 |
| rm_work | if that's wrong, i'll need to adjust in the listener one | 21:29 |
| rm_work | and it's easier to just look at the logic in that patch since it's less busy | 21:29 |
| reedip | anywhere i can help? just got up | 21:30 |
| rm_work | reedip: if you want to test the listeners patch :P | 21:31 |
| rm_work | just pushed an update to it | 21:31 |
| reedip | 424744 ? | 21:31 |
| reedip | rm_work ^^ ? | 21:31 |
| rm_work | yes | 21:31 |
| reedip | k ... | 21:31 |
| reedip | sure .. | 21:32 |
| johnsom | rm_work we need to remember to fix all of these for project ID. Like get_one doesn't care at the moment. Again, I guess that could be a hook | 21:34 |
| johnsom | But would be more efficient as just code to filter the get call | 21:34 |
| rm_work | hmm | 21:36 |
| rm_work | yeah so i was trying to figure out how exactly this interfaces with the other filtering code that's coming down the pipe | 21:36 |
| rm_work | what do you think of the way it is done in my patch... should i start over? | 21:37 |
| rm_work | not sure if that's the right approach or not | 21:37 |
| johnsom | I think it works. I'm just looking at the test now | 21:37 |
| johnsom | Yeah, I am fine with that | 21:38 |
| johnsom | https://www.irccloud.com/pastebin/9eBbDTCW/ | 21:39 |
| rm_work | grrr why did sonar fail | 21:39 |
| rm_work | looks like a legit error but i thought i ran etsts | 21:39 |
| rm_work | ah | 21:39 |
| rm_work | augh | 21:39 |
| rm_work | how | 21:39 |
| rm_work | >_< | 21:39 |
| rm_work | dumb k | 21:40 |
| rm_work | gives me a chance to fix something else I wanted to tweak anyway | 21:40 |
| rm_work | how did this slip through | 21:41 |
| rm_work | i ran functional locally >_< | 21:41 |
| johnsom | Maybe because it's like almost 7am there? | 21:42 |
| rm_work | wait no | 21:42 |
| rm_work | this IS right | 21:42 |
| rm_work | self.conf has a subobject conf | 21:42 |
| rm_work | >_> | 21:42 |
| rm_work | the sonar fail must be something else? | 21:42 |
| johnsom | It failed on zuul too | 21:43 |
| rm_work | aaah i see | 21:43 |
| rm_work | because i had changed base.py | 21:43 |
| rm_work | but i didn't move that over | 21:43 |
| xgerman | I guess somebody gets to eat a hat | 21:43 |
| rm_work | i'll just rebase this on top of the listeners change | 21:43 |
| rm_work | xgerman: yeah it's all prepared and ready to ingest | 21:44 |
| xgerman | well, I will be back in a couple of hours if you need +2s | 21:44 |
| xgerman | (and it tests well) | 21:44 |
| johnsom | Ha, it's pretty darn close. Could still happen pacific time | 21:44 |
| xgerman | ;-) | 21:45 |
| xgerman | ping me if you. need me should be back on 2-3 hours for a bit | 21:45 |
| johnsom | Ok | 21:45 |
| rm_work | ok fixed | 21:45 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix get_all method for v2 LB controller https://review.openstack.org/449822 | 21:45 |
| rm_work | anywho | 21:45 |
| rm_work | listeners should be good | 21:45 |
| rm_work | right now IMO, or close | 21:45 |
| rm_work | unless i missed something else major | 21:45 |
| rm_work | oh i didn't look at the duplicate thing yet | 21:45 |
| rm_work | >_> | 21:45 |
| rm_work | you just created a second listener with the same stuff? | 21:46 |
| johnsom | YEs | 21:46 |
| johnsom | It is odd, the DB should be throwing a fit for that | 21:46 |
| rm_work | prolly something simple | 21:46 |
| rm_work | hmm | 21:48 |
| rm_work | webtest.app.AppError: Bad response: 409 Conflict (not 201) | 21:48 |
| rm_work | {"debuginfo": null, "faultcode": "Client", "faultstring": "Another Listener on this Load Balancer is already using protocol_port 80"} | 21:48 |
| rm_work | that's what I get | 21:48 |
| rm_work | when I try to create the same listener again | 21:48 |
| rm_work | what EXACTLY did you do? | 21:48 |
| johnsom | That is what I would have expected too | 21:48 |
| johnsom | curl -v -X POST -H "Content-Type: application/json" -H "Accept: application/json" -H "X-Auth-Token: $current_token" -d "{\"listener\": {\"protocol\": \"HTTP\", \"description\": \"test description\", \"admin_state_up\": false, \"connection_limit\": 200, \"protocol_port\": \"80\", \"loadbalancer_id\": \"$test_lb_id\", \"name\": \"test-listener\"}}" | 21:49 |
| johnsom | http://$test_API_IP:$test_API_PORT/v2.0/lbaas/listeners | 21:49 |
| johnsom | I ran that twice | 21:49 |
| johnsom | Let me see if I can repro | 21:50 |
| johnsom | Yeah, reproduces for me | 21:51 |
| johnsom | Though I am still a few patches back | 21:52 |
| rm_work | hmmmmmm | 21:52 |
| rm_work | it shouldn't work :/ | 21:53 |
| rm_work | i think I did fix that but it was a WHILE ago | 21:53 |
| rm_work | how far back are you? | 21:53 |
| rm_work | well anyway, I added a test for it | 21:53 |
| rm_work | if I need to make another patchset | 21:53 |
| johnsom | 36bc31b6380557259e2afeda4402efdb4de4c393 I think | 21:53 |
| johnsom | Yeah, I still get it with the current patch. I know I'm running the new code as I have the "" for name now instead of null | 21:57 |
| rm_work | that's only one back | 21:58 |
| rm_work | hmmmmmm | 21:58 |
| rm_work | wtf | 21:58 |
| rm_work | i can't repro | 21:58 |
| rm_work | well | 21:58 |
| rm_work | i can't repro *in functional testing* | 21:58 |
| rm_work | I am not testing a deploy | 21:58 |
| johnsom | Maybe I just have my db scrambled, let me look | 21:58 |
| johnsom | Well, after I look at the changes I will restack and try again | 21:59 |
| rm_work | hmm k | 22:03 |
| rm_work | i'll push up the test | 22:03 |
| rm_work | may as well | 22:03 |
| johnsom | Ok, I'm stacking. I needed barbican anyway | 22:03 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Octavia v2 API for listeners https://review.openstack.org/424744 | 22:04 |
| rm_work | just one functional test ^^ | 22:04 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Implement sorting and pagination for octavia https://review.openstack.org/382147 | 22:06 |
| rm_work | rebase ^^ | 22:07 |
| rm_work | actually... i should rebase that on the end of the chain | 22:07 |
| rm_work | ok some more fixes incoming... | 22:19 |
| rm_work | but not to Listener | 22:20 |
| rm_work | so whatev | 22:20 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for Members https://review.openstack.org/405932 | 22:20 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Implement sorting and pagination for octavia https://review.openstack.org/382147 | 22:20 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Introduce Octavia v2 API for pools https://review.openstack.org/405922 | 22:20 |
| rm_work | alright i am out for a little | 22:20 |
| rm_work | I probably shouldn't +2 anything in this chain at this point anyway | 22:21 |
| johnsom | Ok, have a good one | 22:22 |
| *** sanfern has quit IRC | 22:23 | |
| *** sanfern has joined #openstack-lbaas | 22:26 | |
| *** eezhova has quit IRC | 22:28 | |
| johnsom | Yeah, still reproducing that bug with duplicate listeners | 22:36 |
| *** sanfern has quit IRC | 22:39 | |
| rm_work | wtf | 22:39 |
| rm_work | how | 22:39 |
| *** fnaval has quit IRC | 22:39 | |
| rm_work | hold on, what was the trace again | 22:39 |
| johnsom | Not sure, debugging now | 22:40 |
| johnsom | https://www.irccloud.com/pastebin/CjjmehSQ/ | 22:40 |
| rm_work | yeah it should get caught in except odb_exceptions.DBDuplicateEntry | 22:41 |
| rm_work | aaah | 22:44 |
| rm_work | it's a multi-column key isn't it | 22:44 |
| rm_work | if ['id'] == de.columns | 22:44 |
| rm_work | or ... no | 22:44 |
| rm_work | it should just be ID | 22:44 |
| johnsom | 2017-03-24 15:42:40.301 120351 ERROR octavia.api.v2.controllers.listener [req-44cfd488-b2f3-4b86-b1ff-428135404e43 - bfd03e05311d4fafa659339cb3d84c60 - default default] {'insert_headers': None, 'l7policies': [], 'protocol': 'HTTP', 'description': 'test description', 'default_pool': None, 'enabled': False, 'provisioning_status': 'PENDING_CREATE', | 22:45 |
| johnsom | 'connection_limit': 200, 'id': '2292d5dd-e9de-4c03-8556-0d3f860c78f4', 'tls_certificate_id': None, 'project_id': u'bfd03e05311d4fafa659339cb3d84c60', 'protocol_port': 80, 'load_balancer_id': u'8832d83a-dc96-42f7-b879-da3472980db9', 'default_pool_id': None, 'operating_status': 'OFFLINE', 'name': 'test-listener'} | 22:45 |
| johnsom | Some debug logging of the listener_dict right before the create | 22:45 |
| rm_work | uhh | 22:45 |
| johnsom | The constraint that should fire is UNIQUE KEY `uq_listener_load_balancer_id_protocol_port` (`load_balancer_id`,`protocol_port`), | 22:46 |
| *** sanfern has joined #openstack-lbaas | 22:46 | |
| rm_work | yeah | 22:47 |
| rm_work | just saw that | 22:47 |
| rm_work | i'm tired T_T | 22:47 |
| rm_work | but yeah it should be hitting that | 22:47 |
| johnsom | The other odd thing, it's not creating a record | 22:48 |
| rm_work | ummmmm | 22:48 |
| rm_work | like | 22:48 |
| rm_work | the first one isn't created? | 22:48 |
| rm_work | or, the second one isn't | 22:48 |
| johnsom | Ok, so odd, the first exception is firing | 22:49 |
| *** fnaval has joined #openstack-lbaas | 22:49 | |
| johnsom | It's walking off the end of the except block | 22:51 |
| rm_work | O_o | 22:51 |
| johnsom | de.columns == ['uq_listener_load_balancer_id_protocol_port'] | 22:52 |
| johnsom | Well, that is interesting | 22:52 |
| rm_work | uhh | 22:53 |
| rm_work | it's the unique constraint NAME and not the columns?? | 22:53 |
| rm_work | sqlalchemy change? | 22:53 |
| rm_work | let me rebuild my env | 22:53 |
| johnsom | I did this: LOG.error(de.columns) | 22:54 |
| johnsom | 2017-03-24 15:52:01.362 120596 ERROR octavia.api.v2.controllers.listener [req-02e43177-7b30-4a4c-bd97-7096b94159e4 - bfd03e05311d4fafa659339cb3d84c60 - default default] ['uq_listener_load_balancer_id_protocol_port'] | 22:54 |
| rm_work | ok so i'm on SQLAlchemy 1.0.17 | 22:54 |
| rm_work | updating | 22:54 |
| johnsom | SQLAlchemy (1.0.17) | 22:55 |
| rm_work | ... | 22:55 |
| johnsom | mysql-server 5.7.17-0ubuntu0.16.04.1 | 22:55 |
| rm_work | this is sqlite | 22:55 |
| rm_work | sooooo | 22:55 |
| rm_work | i could see it being different, because it always is | 22:55 |
| johnsom | Yeah, I'm running live | 22:56 |
| johnsom | so It's mysql | 22:56 |
| rm_work | so guessing sqlite returns column names, mysql returns constraint name | 22:56 |
| rm_work | even though it's called .columns | 22:57 |
| johnsom | That would be my guess as well.... Another sqlalchemy failure | 22:57 |
| rm_work | uhh | 22:57 |
| rm_work | maybe we need to update | 22:57 |
| rm_work | can you manually update sqlalchemy | 22:57 |
| rm_work | in your devstack | 22:57 |
| rm_work | it has like 1.1.6 | 22:57 |
| rm_work | maybe they already fixed it | 22:57 |
| johnsom | Nope, same thing with 1.1.6 | 22:59 |
| johnsom | You have much higher expectations of sqlalchemy than my experience leads me to have | 22:59 |
| *** aojea has quit IRC | 22:59 | |
| johnsom | Well, this is easy enough to fix | 22:59 |
| johnsom | I'm going to push up a fix | 23:01 |
| rm_work | err | 23:05 |
| rm_work | ok but | 23:05 |
| rm_work | just doing an or? | 23:05 |
| johnsom | yeah | 23:06 |
| rm_work | and no way to test for it <_< | 23:06 |
| johnsom | The joy of sqlite. I mean we could switch functional over to running with a mysql DB... It's some project-config fun | 23:07 |
| rm_work | errr | 23:08 |
| rm_work | that'd be difficult because functionals run locally | 23:08 |
| rm_work | and ... after that, they wouldn't | 23:08 |
| rm_work | <_< | 23:08 |
| johnsom | Right, | 23:09 |
| *** ndahiwade has joined #openstack-lbaas | 23:10 | |
| johnsom | I am now going to spend the reset of the day trying to get pep8 happy with these, ugh.... | 23:13 |
| rm_work | ? | 23:14 |
| johnsom | long strings, etc. Getting the wrapping right | 23:15 |
| rm_work | it COULD be an oslo_db bug | 23:17 |
| rm_work | actually | 23:17 |
| rm_work | technically | 23:17 |
| rm_work | probably not <_< | 23:17 |
| rm_work | k, really out | 23:19 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Octavia v2 API for listeners https://review.openstack.org/424744 | 23:19 |
| *** links has quit IRC | 23:24 | |
| ankur-gupta-f4 | johnsom: https://review.openstack.org/#/c/403938/ is it even necessary to add the apiv1 testenv if I already discussed how to run the scenario tests? | 23:39 |
| johnsom | Two different tempest test suites | 23:40 |
| johnsom | In theory | 23:40 |
| ankur-gupta-f4 | So what would i write | 23:41 |
| johnsom | But, yeah, that could be too much detail. It's not like you are listing every test option here | 23:41 |
| ankur-gupta-f4 | the apiv1 testenv runs the tempest api tests and the scenario runs scenario tests | 23:41 |
| ankur-gupta-f4 | that is just confusing | 23:41 |
| johnsom | Sorry, slow typing, my main system is patching... | 23:42 |
| ankur-gupta-f4 | how would one use the --failing and --until-failure flags | 23:48 |
| openstackgerrit | Ankur proposed openstack/octavia master: Introduce TESTING.rst for quick developer reference https://review.openstack.org/403938 | 23:52 |
| johnsom | Tox -e py27 -- --failing | 23:52 |
| johnsom | It only runs the tests that failed the last run | 23:53 |
| johnsom | Until fail runs the tests in a loop, randomizing the order each time (as it always does) | 23:53 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!