| *** goldyfruit_ has joined #openstack-lbaas | 00:12 | |
| johnsom | rm_work Still around? | 00:17 |
|---|---|---|
| rm_work | yeah | 00:17 |
| rm_work | in a meeting but yes | 00:17 |
| johnsom | https://review.opendev.org/#/c/659626/15/octavia/db/models.py Line 810 | 00:18 |
| johnsom | Carlos has a foreignkey to the listeners table for listener_id. When you get the sqlalchemy model back it automagicly has the whole listener object also attached. Do you know how to stop that? | 00:18 |
| johnsom | a dir(db_object) shows: | 00:20 |
| johnsom | ['__class__', '__delattr__', '__dict__', '__doc__', '__eq__', '__format__', '__getattribute__', '__hash__', '__init__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_find_in_graph', '_get_unique_key', '_name', 'cidr', 'from_dict', 'listener', 'listener_id', 'to_dict', 'update'] | 00:20 |
| rm_work | uhh | 00:20 |
| johnsom | My problem is our wonderful to_dict() will expand that listener object too. | 00:20 |
| rm_work | if it's lazy-load, then it won't until you access it | 00:20 |
| rm_work | ah hmm | 00:21 |
| rm_work | yeah that'd access it, lol | 00:21 |
| rm_work | i am not sure | 00:21 |
| johnsom | Ok, thanks | 00:21 |
| johnsom | I don't want to do something dumb like "if listenercidr del listener" | 00:21 |
| johnsom | lol, well, this will work | 00:31 |
| johnsom | def to_dict(self, **kwargs): | 00:31 |
| johnsom | return {'cidr': self.cidr, 'listener_id': self.listener_id} | 00:31 |
| rm_work | :D | 00:33 |
| johnsom | The things I do to work around sqlalchemy issues..... | 00:36 |
| *** henriqueof has joined #openstack-lbaas | 00:45 | |
| *** henriqueof1 has quit IRC | 00:46 | |
| *** ivve has quit IRC | 00:47 | |
| *** ricolin has joined #openstack-lbaas | 00:48 | |
| *** henriqueof1 has joined #openstack-lbaas | 00:56 | |
| *** henriqueof has quit IRC | 00:57 | |
| *** happyhemant has quit IRC | 01:00 | |
| *** henriqueof1 has quit IRC | 01:05 | |
| *** henriqueof has joined #openstack-lbaas | 01:05 | |
| *** pcaruana has quit IRC | 01:09 | |
| johnsom | rm_work are you working on the nits patch? | 02:03 |
| rm_work | waiting for the recheck | 02:04 |
| johnsom | Oh it exploded in zuul | 02:04 |
| johnsom | If you have that I will finish up fixing Carlos's patch | 02:04 |
| rm_work | yeah | 02:09 |
| johnsom | I would just ignore W504 | 02:14 |
| rm_work | err, where? | 02:15 |
| rm_work | i didn't get that one | 02:15 |
| johnsom | https://github.com/openstack/octavia/blob/master/tox.ini#L144 | 02:16 |
| johnsom | Well it is a bit of a stupid rule, so I would just exclude it. At least for now | 02:16 |
| johnsom | Yeah, neutron ignores it too | 02:17 |
| rm_work | i mean, which patch is it causing problems? | 02:51 |
| rm_work | oh weird | 02:52 |
| rm_work | i don't get that Warning locally | 02:53 |
| rm_work | can't repro | 02:53 |
| rm_work | that or E305 | 02:53 |
| rm_work | ok well I think I got those handled | 03:00 |
| rm_work | will try again in the gate since I can't repro it locally T_T | 03:01 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 03:01 |
| johnsom | Did you rebuild your tox environment? It shows up for me | 03:02 |
| rm_work | yeah, like 3 times just now | 03:03 |
| johnsom | Super strange | 03:03 |
| rm_work | hmmm locally it's still installing 2.5.5 >_> | 03:04 |
| rm_work | wtf | 03:04 |
| rm_work | weird | 03:05 |
| rm_work | something cached? <_< | 03:05 |
| sapd1 | rm_work, you are working on VIP ACL? | 03:06 |
| johnsom | sapd1: I am | 03:06 |
| sapd1 | Do we need to write specs about the last discussion about deny access for LB? | 03:07 |
| sapd1 | Just write story on storyboard, rights? | 03:07 |
| johnsom | I think so, but Mr PTL can chime in | 03:07 |
| johnsom | A spec might be good as it allows comments better than storyboard | 03:10 |
| rm_work | I'm concerned though, because this is ... comments on the way we're implementing... the thing that we're planning to merge in the next 24 hours? lol | 03:11 |
| johnsom | Well, I think the proposal is the inverse, but yeah, this needs discussion given the decision we made for this patch. | 03:13 |
| rm_work | johnsom: i am a little concerned that there's broken reqs in upper-requirements: https://zuul.opendev.org/t/openstack/build/9782d30ea93e46f8b53b82df302003dd | 03:14 |
| rm_work | for py2 | 03:14 |
| rm_work | everything py2 is failing now | 03:14 |
| rm_work | py27, functional, py2-scenarios ... | 03:15 |
| johnsom | Yeah, I saw that. I wasn’t sure if it was something we were doing that caused that or not | 03:15 |
| rm_work | i don't think so | 03:16 |
| johnsom | Do we just need to advance something in our requirements? | 03:16 |
| rm_work | it's an == coming from upper-constraints | 03:16 |
| rm_work | which is from requirements project, isn't it? | 03:16 |
| johnsom | Yes | 03:17 |
| rm_work | ERROR: No matching distribution found for configparser===4.0.1 (from -c /opt/upper-constraints.txt (line 620)) | 03:17 |
| rm_work | so ... | 03:17 |
| rm_work | I think maybe they f'd up a patch? | 03:17 |
| rm_work | checking | 03:17 |
| rm_work | yeah the last patch broke it | 03:18 |
| rm_work | https://github.com/openstack/requirements/commit/c9b1610a12b4a75e2e658dd85ce2650ab358b60f#diff-0bdd949ed8a7fdd4f95240bd951779c8L620 | 03:18 |
| johnsom | We don’t have that in our reqs.... sigh | 03:18 |
| sapd1 | https://storyboard.openstack.org/#!/story/2006533 have a look. | 03:18 |
| rm_work | i don't think we require it directly? | 03:19 |
| rm_work | we never use it directly, it comes in from somewhere else | 03:19 |
| rm_work | so we shouldn't have it, i think | 03:19 |
| rm_work | but... there is no 4.0.1 | 03:19 |
| rm_work | https://pypi.org/project/configparser/#history | 03:19 |
| johnsom | Lol, wow, how did that happen. Thode! | 03:20 |
| rm_work | poking him >_> | 03:21 |
| rm_work | ONE THING AFTER ANOTHER | 03:21 |
| rm_work | love freeze week :D | 03:21 |
| rm_work | yeah I am hereby stating that we're entering freeze next week instead. fight me. | 03:21 |
| johnsom | Lol | 03:22 |
| rm_work | Friday, September 20 | 03:22 |
| * johnsom thinks he needs to move his vacation up two weeks next year | 03:22 | |
| rm_work | I think that works | 03:22 |
| rm_work | eh, maybe Wednesday is enough | 03:23 |
| rm_work | gives us the weekend for stuff to settle and a couple days to get these things merged | 03:23 |
| rm_work | Wednesday the 18th? That seem good to you? | 03:23 |
| johnsom | Yeah, we don’t do milestone releases anymore so.. | 03:23 |
| johnsom | PTL call | 03:24 |
| rm_work | hmm, what to do about: W605 invalid escape sequence '\ ' | 03:30 |
| rm_work | add another `\`? | 03:31 |
| rm_work | that was your code | 03:32 |
| rm_work | right johnsom ? | 03:32 |
| rm_work | do you know what it's supposed to be doing? is there a reason it's not a real escape? | 03:32 |
| johnsom | ummm, link? | 03:33 |
| rm_work | ah right | 03:33 |
| rm_work | https://github.com/openstack/octavia/blob/master/octavia/common/jinja/haproxy/combined_listeners/jinja_cfg.py#L143 | 03:34 |
| rm_work | there's one more but it's exactly the same, so whatever we do for this one | 03:34 |
| rm_work | in my head i already showed you the code, lol | 03:34 |
| johnsom | Ah, it is supposed to be an explicit \ + space, so in the string I'm escaping the space. Not for python, but for the output | 03:35 |
| rm_work | yeah ok | 03:35 |
| johnsom | so maybe just r'\ '? | 03:35 |
| rm_work | so we need to ignore that error | 03:35 |
| rm_work | ah err | 03:35 |
| rm_work | so yeah \\ would also work | 03:35 |
| rm_work | that's what that means | 03:35 |
| johnsom | Yeah, r is bad, don't do that | 03:36 |
| johnsom | oh, well, maybe r is ok. | 03:36 |
| rm_work | it should be | 03:37 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 03:37 |
| rm_work | but i'll just explicitly escape it | 03:37 |
| johnsom | Yeah, ok, it should just be r'\ ' | 03:37 |
| rm_work | '\\ ' is == r'\ ' | 03:37 |
| johnsom | I will quote you on that | 03:37 |
| johnsom | grin | 03:37 |
| rm_work | >>> '\\ ' is r'\ ' | 03:37 |
| rm_work | True | 03:37 |
| rm_work | i mean it quite literally lol | 03:38 |
| johnsom | In python 99.3? | 03:38 |
| johnsom | revenge of the python? | 03:38 |
| rm_work | uhhh we'll see in a decade or five i guess | 03:38 |
| johnsom | lol | 03:38 |
| rm_work | alright we're gonna be queued for another 1-2h i guess so | 03:39 |
| rm_work | i'm gonna go do stuff | 03:39 |
| rm_work | be back in 2--6h | 03:39 |
| rm_work | *2-6h | 03:39 |
| johnsom | o/ | 03:39 |
| johnsom | I am still working on straightening out the cidr object stuff. Solved one problem, found another one | 03:40 |
| *** KeithMnemonic has quit IRC | 04:04 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add VIP access control list https://review.opendev.org/659626 | 04:06 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add VIP access control list https://review.opendev.org/659626 | 04:08 |
| johnsom | Ok, I think that is working. | 04:08 |
| johnsom | Pretty pumped that my new lib-getter functional test caught some issues. | 04:09 |
| johnsom | Ok, calling it a night. I think we are stuck until requirements can fix upper-constraints.... | 04:11 |
| *** spatel has joined #openstack-lbaas | 04:17 | |
| *** sapd1_x has joined #openstack-lbaas | 04:23 | |
| *** ricolin has quit IRC | 04:27 | |
| *** pcaruana has joined #openstack-lbaas | 04:32 | |
| *** pcaruana has quit IRC | 04:58 | |
| *** spatel has quit IRC | 04:58 | |
| *** sapd1_x has quit IRC | 05:13 | |
| *** vishalmanchanda has joined #openstack-lbaas | 05:20 | |
| rm_work | Yeah | 05:29 |
| *** ramishra has joined #openstack-lbaas | 05:31 | |
| *** gcheresh_ has joined #openstack-lbaas | 05:42 | |
| *** tkajinam has quit IRC | 06:00 | |
| *** tkajinam has joined #openstack-lbaas | 06:06 | |
| *** pcaruana has joined #openstack-lbaas | 06:08 | |
| *** rpittau|afk is now known as rpittau | 06:36 | |
| *** pcaruana has quit IRC | 06:41 | |
| *** ricolin has joined #openstack-lbaas | 06:41 | |
| *** maciejjozefczyk has joined #openstack-lbaas | 06:43 | |
| *** tkajinam_ has joined #openstack-lbaas | 06:53 | |
| *** tkajinam has quit IRC | 06:56 | |
| *** ricolin has quit IRC | 06:59 | |
| *** ccamposr__ has joined #openstack-lbaas | 07:00 | |
| *** ccamposr has quit IRC | 07:03 | |
| *** ivve has joined #openstack-lbaas | 07:07 | |
| cgoncalves | johnsom, thanks for the great help on the VIP ACL patch! | 07:08 |
| cgoncalves | tox started now complaining about E741 ambiguous variable name 'l'... | 07:09 |
| cgoncalves | sorry, pep8 | 07:09 |
| cgoncalves | the exact line has been there for a while, so maybe yet-another requirement bump? :S | 07:10 |
| *** pcaruana has joined #openstack-lbaas | 07:11 | |
| *** trident has quit IRC | 07:15 | |
| cgoncalves | pep8 passed locally, though | 07:15 |
| *** ricolin has joined #openstack-lbaas | 07:18 | |
| gthiemonge | cgoncalves: you should try it with 'tox --recreate' | 07:21 |
| *** happyhemant has joined #openstack-lbaas | 07:23 | |
| cgoncalves | gthiemonge, I did http://paste.openstack.org/show/kgwvKXsiyDHWBSRohyjP/ | 07:24 |
| *** trident has joined #openstack-lbaas | 07:24 | |
| cgoncalves | changed basepython to python2 in tox.ini. it failed now but different problem | 07:24 |
| *** bcafarel has quit IRC | 07:24 | |
| gthiemonge | hm, I have hundreds of errors | 07:27 |
| *** trident has quit IRC | 07:28 | |
| gthiemonge | cgoncalves: http://paste.openstack.org/show/775346/ | 07:28 |
| cgoncalves | gthiemonge, reading channel backlog, johnsom is facing same problem as you while rm_work could not reproduce it | 07:31 |
| cgoncalves | re-recreated a second time, passed | 07:31 |
| cgoncalves | configparser update reverted https://review.opendev.org/#/c/681630/ | 07:32 |
| johnsom | Yeah, it is a requirements team issue as far as I know | 07:34 |
| cgoncalves | lol, autobahn lib (https://pypi.org/project/autobahn/). the autobahn here in Germany has been in intense refactor mode since I moved to Germany 5 years ago | 07:35 |
| *** trident has joined #openstack-lbaas | 07:37 | |
| johnsom | I just remember white knuckles as the BMW flew by is. | 07:44 |
| johnsom | Ok, offline for the night. Good luck | 07:44 |
| *** bcafarel has joined #openstack-lbaas | 07:56 | |
| rm_work | cgoncalves: all of those errors are fixed in the patch I submitted above | 07:58 |
| rm_work | https://review.opendev.org/681144 | 07:58 |
| rm_work | We needed the reqs patch to land before we can merge that one | 07:58 |
| rm_work | But that will fix our gate | 07:58 |
| rm_work | I recommend you base all other patches on top of that for now | 07:59 |
| rm_work | As it will be the first to merge once reqs are working | 07:59 |
| rm_work | gthiemonge: ^^ | 07:59 |
| cgoncalves | py27 is passing now in Zuul | 07:59 |
| cgoncalves | rebasing now. thanks, rm_work! | 08:00 |
| gthiemonge | rm_work: ack, thanks | 08:00 |
| cgoncalves | oh, it is already. thank you, johnsom! | 08:00 |
| cgoncalves | rm_work, you missed v1/tasks/database_tasks.py | 08:01 |
| cgoncalves | ./octavia/controller/worker/v1/tasks/database_tasks.py:1590:13: E741 ambiguous variable name 'l' | 08:01 |
| rm_work | Thought I got that one | 08:02 |
| rm_work | I remember fixing it... | 08:02 |
| rm_work | Oh ffff yeah I did it once not twice :D | 08:02 |
| rm_work | Ooos | 08:02 |
| rm_work | *ooos | 08:02 |
| rm_work | *oops | 08:03 |
| rm_work | Stupid phone | 08:03 |
| rm_work | Sec I'll fix the other one | 08:03 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 08:05 |
| *** tkajinam_ has quit IRC | 08:09 | |
| rm_work | if you want to +2 it I will +A :D | 08:10 |
| cgoncalves | waiting for Zuul to pass certain jobs | 08:13 |
| rm_work | yeah ok, i'll +2 and you can +A when you're happy :) | 08:16 |
| rm_work | cause i'm about to head to bed | 08:16 |
| *** threestrands has quit IRC | 08:16 | |
| cgoncalves | k | 08:16 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Fix 'additive_only' parameter api-ref https://review.opendev.org/681625 | 08:16 |
| rm_work | rebased and +A | 08:16 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Add VIP access control list https://review.opendev.org/659626 | 08:18 |
| rm_work | WAT | 08:28 |
| rm_work | octavia-tox-functional-py37-tips FAILED | 08:28 |
| rm_work | ERROR: Could not find a version that satisfies the requirement SQLAlchemy===1.3.8 | 08:28 |
| rm_work | I will *destroy* someone | 08:29 |
| rm_work | ok, no, that exists in pypi | 08:29 |
| rm_work | wtf | 08:29 |
| rm_work | ah, connection issue | 08:30 |
| rm_work | ok, well, that'll need a recheck -- cgoncalves can you do that when it finishes? | 08:30 |
| cgoncalves | rm_work, we can force a recheck with a rebase or something | 08:30 |
| rm_work | nothing to rebase it on | 08:30 |
| rm_work | i was thinking that lol | 08:31 |
| rm_work | could make a change to the commit message <_< | 08:31 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 08:31 |
| cgoncalves | rm_work, ^ | 08:31 |
| rm_work | +2 | 08:32 |
| *** bcafarel has quit IRC | 08:34 | |
| *** bcafarel has joined #openstack-lbaas | 08:35 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia stable/stein: Add support for monitor_{address,port} in UDP members https://review.opendev.org/675349 | 08:43 |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia stable/stein: Fixed pool and members status with UDP loadbalancers https://review.opendev.org/675362 | 08:43 |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia stable/stein: Fixed down server issue after reloading keepalived https://review.opendev.org/675363 | 08:43 |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia stable/stein: Prevent UDP LBs to use different IP protocol versions in amphora driver https://review.opendev.org/675364 | 08:43 |
| *** sapd1 has quit IRC | 08:44 | |
| *** ccamposr has joined #openstack-lbaas | 08:48 | |
| *** ccamposr__ has quit IRC | 08:51 | |
| *** ccamposr__ has joined #openstack-lbaas | 09:02 | |
| *** ccamposr has quit IRC | 09:04 | |
| cgoncalves | Err:4 http://mirror.regionone.fortnebula.opendev.org/ubuntu bionic-security InRelease | 09:10 |
| cgoncalves | now on octavia-v2-dsvm-scenario :/ | 09:10 |
| cgoncalves | fortnebula has caused nothing by trouble | 09:11 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 09:11 |
| cgoncalves | uhm, +2s are cleared but +1s aren't?! | 09:12 |
| cgoncalves | more mirror issues with fortnebula | 09:13 |
| cgoncalves | Unable to connect to mirror.regionone.fortnebula.opendev.org | 09:13 |
| *** rcernin has quit IRC | 09:16 | |
| *** ricolin has quit IRC | 09:37 | |
| cgoncalves | fortnebula strikes again... | 10:06 |
| cgoncalves | going to -infra | 10:07 |
| cgoncalves | Disable fortnebula cloud - https://review.opendev.org/#/c/681697/ | 10:11 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Convert listener flows to use provider models https://review.opendev.org/660236 | 10:27 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Convert pool flows to use dicts https://review.opendev.org/665381 | 10:27 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Transition member flows to use dicts https://review.opendev.org/657842 | 10:27 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Transition amphora flows to dicts https://review.opendev.org/668898 | 10:27 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Convert Lb flows to use provider dicts https://review.opendev.org/671725 | 10:27 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: [WIP] Jobboard based controller https://review.opendev.org/647406 | 10:27 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 10:32 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Convert health_monitor flows to use provider models https://review.opendev.org/679779 | 10:42 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Transition l7policy flows to dicts https://review.opendev.org/665977 | 10:47 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Transition l7rule flows to dicts https://review.opendev.org/668173 | 10:47 |
| *** ajay33 has joined #openstack-lbaas | 10:48 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Use retry for AmphoraComputeConnectivityWait https://review.opendev.org/662791 | 10:51 |
| *** sapd1_x has joined #openstack-lbaas | 10:56 | |
| *** numans has joined #openstack-lbaas | 10:58 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Convert Lb flows to use provider dicts https://review.opendev.org/671725 | 11:24 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: [WIP] Jobboard based controller https://review.opendev.org/647406 | 11:24 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Add VIP access control list https://review.opendev.org/659626 | 11:41 |
| *** boden has joined #openstack-lbaas | 11:47 | |
| *** ccamposr__ has quit IRC | 12:12 | |
| *** ccamposr has joined #openstack-lbaas | 12:13 | |
| *** goldyfruit_ has quit IRC | 12:16 | |
| *** pcaruana has quit IRC | 13:30 | |
| *** ricolin has joined #openstack-lbaas | 13:39 | |
| *** Vorrtex has joined #openstack-lbaas | 13:41 | |
| *** happyhemant has quit IRC | 13:47 | |
| *** KeithMnemonic has joined #openstack-lbaas | 13:53 | |
| *** goldyfruit_ has joined #openstack-lbaas | 13:53 | |
| *** goldyfruit_ has quit IRC | 13:54 | |
| *** tkajinam has joined #openstack-lbaas | 13:55 | |
| *** goldyfruit has joined #openstack-lbaas | 13:56 | |
| openstackgerrit | Maciej Józefczyk proposed openstack/octavia master: Add new algorithm SOURCE_IP_PORT https://review.opendev.org/672463 | 14:00 |
| openstackgerrit | Maciej Józefczyk proposed openstack/octavia master: Validate supported LB algorithm in Amphora provider drivers https://review.opendev.org/672477 | 14:00 |
| *** pcaruana has joined #openstack-lbaas | 14:08 | |
| *** goldyfruit_ has joined #openstack-lbaas | 14:09 | |
| *** goldyfruit has quit IRC | 14:12 | |
| *** henriqueof1 has joined #openstack-lbaas | 14:27 | |
| *** henriqueof has quit IRC | 14:30 | |
| *** gcheresh_ has quit IRC | 14:40 | |
| *** Vorrtex has quit IRC | 14:51 | |
| *** pcaruana has quit IRC | 14:57 | |
| *** tkajinam has quit IRC | 15:06 | |
| *** rouk has joined #openstack-lbaas | 15:11 | |
| rouk | johnsom: in ProviderDriver, create_vip_port has a message of 'This provider does not support creating VIP ports. Octavia will create it.', but it crashes and gives up if that function isnt in the driver. | 15:13 |
| cgoncalves | johnsom, rm_work: the ACL patch is passing now. could you +2 it again? https://review.opendev.org/#/c/659626/ | 15:13 |
| rouk | is it supposed to just die? should it be catching and continuing with creation itself? | 15:14 |
| cgoncalves | I thought it had been fixed. IIRC problem was we were catching an exception from octavia whereas it should be the same but from octavia-lib (or the other way around) | 15:15 |
| johnsom | rouk Umm, no, that should not happen. What version of Octavia and what driver are you using? | 15:15 |
| johnsom | Yeah, I also thought that was fixed | 15:15 |
| rouk | stein, and this a10 one im trying to make work. | 15:16 |
| rouk | but i think it might be an error directly after with an empty message thats actually stopping it. | 15:16 |
| rouk | "ProviderDriverError: Provider 'a10' reports error: " no message. | 15:16 |
| johnsom | Can you paste.openstack.org the log sections? | 15:17 |
| rouk | yep, sec | 15:17 |
| rouk | http://paste.openstack.org/show/5IOiitE3yBDUbEmMbdLN/ | 15:18 |
| rouk | https://github.com/a10networks/a10-octavia/blob/master/a10_octavia/api/drivers/driver.py | 15:19 |
| rouk | the driver is here. | 15:19 |
| rouk | i was demo'd this stuff "working", but they must have had some strange undocumented setup. | 15:20 |
| johnsom | lol | 15:21 |
| johnsom | rouk Ok, yeah, we fixed this on master, but forgot to backport it to stein. Sorry about that | 15:26 |
| rouk | oi | 15:26 |
| johnsom | This patch: https://github.com/openstack/octavia/commit/e2defa6dd238d82ce4ee457e21ca047888d27106 | 15:27 |
| rouk | should i cherrypick or just pull up to master? master + stein wont break? | 15:28 |
| openstackgerrit | Michael Johnson proposed openstack/octavia stable/stein: Fix catching driver exceptions https://review.opendev.org/681779 | 15:28 |
| rouk | or are you going to merge it to stable/stein in the next 10 mins and ill just rebuild? | 15:28 |
| johnsom | merge in 10 minutes? What mythical CI environment do you work in? grin | 15:29 |
| johnsom | It will be at least a day to merge. More before a release lands on pypi | 15:29 |
| openstackgerrit | Michael Johnson proposed openstack/octavia stable/rocky: Fix catching driver exceptions https://review.opendev.org/681780 | 15:30 |
| cgoncalves | speaking of stable branches, I'll be checking them next week. hopefully we can merge all open stable patches and cut stable releases | 15:30 |
| rouk | am i safer to just run master for my testing? master+stein should be fine for octavia right? | 15:30 |
| johnsom | Yes, it should work fine | 15:31 |
| rouk | alright, ill just build some master octavia images and play with it there. | 15:31 |
| johnsom | We have added a new tag for patches to try to improve the backport situation. I hope there will now be less we miss. | 15:32 |
| johnsom | rouk FYI, you might want to pull master after this next patch merges. There was a requirements issue last night with upper-constraints. | 15:34 |
| johnsom | This patch is just about to merge: https://review.opendev.org/681144 | 15:35 |
| rouk | ah, ETA? | 15:35 |
| johnsom | Any minute now | 15:35 |
| rouk | alright, ill watch for it. | 15:36 |
| *** ataraday_ has quit IRC | 15:36 | |
| *** sapd1_x has quit IRC | 15:41 | |
| openstackgerrit | Merged openstack/octavia master: Fix a few nits from the main volume-based patch https://review.opendev.org/681144 | 15:42 |
| johnsom | rouk ^^^^ Ok, master should be good now. | 15:43 |
| rouk | excellent, and if i just follow requirements.txt for octavia_lib master should be happy with.... 1.3.1? | 15:44 |
| *** rpittau is now known as rpittau|afk | 15:44 | |
| rouk | other way around, words, hard. octavia master is still happy with _lib 1.3.1? | 15:44 |
| johnsom | hmmm, you might need 1.4.0 octavia-lib | 15:45 |
| cgoncalves | yes https://github.com/openstack/octavia/blob/master/lower-constraints.txt#L75 | 15:45 |
| johnsom | Ah, ok, cool | 15:45 |
| *** maciejjozefczyk has quit IRC | 15:45 | |
| rouk | the repo says 1.3.1, just wanted to make sure it was still in sync, since thats the bits im testing against. | 15:46 |
| johnsom | cgoncalves Hmm, should your patch advance that? | 15:46 |
| cgoncalves | johnsom, the allowed_cidrs patch in -lib is part of 1.3.0 or 1.3.1 | 15:46 |
| johnsom | ok | 15:46 |
| cgoncalves | yep, included in 1.3.0 | 15:47 |
| johnsom | Yeah, ok, 1.4.0 was mostly docs and removing the need for oslo log, so not critical | 15:47 |
| *** ivve has quit IRC | 15:52 | |
| *** ricolin has quit IRC | 16:03 | |
| *** ricolin has joined #openstack-lbaas | 16:04 | |
| *** ricolin has quit IRC | 16:05 | |
| *** goldyfruit_ has quit IRC | 16:32 | |
| *** goldyfruit_ has joined #openstack-lbaas | 16:48 | |
| *** Vorrtex has joined #openstack-lbaas | 17:16 | |
| *** ivve has joined #openstack-lbaas | 17:31 | |
| *** pcaruana has joined #openstack-lbaas | 17:58 | |
| rouk | where does line 45 in octavia/controller/queue/v1/endpoints.py get that function from? http://paste.openstack.org/show/woNg9bNPjjXH7uVwcygu/ getting this on my driver lb create. | 18:19 |
| rouk | loadbalancer_create in -lib only takes 2 args in the skeleton, is it getting a different function? | 18:20 |
| johnsom | rouk The queue and endpoints are only used for the amphora driver. | 18:31 |
| rouk | then why are they failing for a not-amphora driver... | 18:31 |
| johnsom | Fair question | 18:31 |
| johnsom | Ah, it looks like the A10 driver is also using a messaging queue. | 18:32 |
| johnsom | rouk Yeah, so, it looks like that A10 driver is putting messages on the Octavia Amphora driver message queue here: https://github.com/a10networks/a10-octavia/blob/master/a10_octavia/api/drivers/driver.py#L54 | 18:35 |
| johnsom | They should use their own queue, not the amphora driver queue | 18:36 |
| johnsom | If you see, they copied the amphora driver queue setup: https://github.com/openstack/octavia/blob/master/octavia/api/drivers/amphora_driver/v1/driver.py#L43 | 18:37 |
| johnsom | Same topic and namespace | 18:37 |
| rouk | they also stole config options, its a mess, but im trying to get it working enough to provide them detailed feedback on the mistakes they need to fix. | 18:40 |
| johnsom | Have them start here: https://docs.openstack.org/octavia/latest/contributor/guides/providers.html grin | 18:41 |
| rouk | yeah... trying to help more than that... i can fix most of it i think pretty quickly. | 18:42 |
| rouk | just not understanding how that function gets pulled so i can find what is wrong. | 18:42 |
| johnsom | Their docs appear to be behind a login, so I can't really give much guidence | 18:43 |
| rouk | the readme on that repo is the only docs you get, this module isnt part of their official support | 18:43 |
| rouk | and their readme is wrong. | 18:43 |
| johnsom | So the driver gets pulled in here: https://github.com/a10networks/a10-octavia/blob/master/setup.cfg#L37 | 18:43 |
| johnsom | So this gets called: https://github.com/a10networks/a10-octavia/blob/master/a10_octavia/api/drivers/driver.py#L51 | 18:44 |
| johnsom | on an LB create | 18:44 |
| johnsom | That is putting a message on our queue which doesn't match the signature for our endpoints (it's an old fork I guess) | 18:44 |
| rouk | but why would it be getting more args? it matches the boilerplate class its extending | 18:45 |
| rouk | also, unrelated question, access logs, is there no way for a user to get the hits from the perspective of the amphora? | 18:46 |
| johnsom | Ours expects three: https://github.com/openstack/octavia/blob/master/octavia/controller/queue/v1/endpoints.py#L42 | 18:46 |
| johnsom | You want this Train feature: https://docs.openstack.org/octavia/latest/admin/log-offloading.html | 18:47 |
| rouk | but the base class is https://github.com/openstack/octavia-lib/blob/master/octavia_lib/api/drivers/provider_base.py#L63 which expects 2 | 18:47 |
| johnsom | Right, because the queue in the Amphora driver is after that interface. | 18:48 |
| johnsom | This is the amphora driver version of the driver interface: https://github.com/openstack/octavia/blob/master/octavia/api/drivers/amphora_driver/v1/driver.py#L71 | 18:49 |
| johnsom | The queue interface is not the driver interface. The queue for the amphora driver is internal to amphora driver. Third party drivers should never use it | 18:50 |
| johnsom | A user call flows like this in Octavia Amphora driver: API->driver interface->amphora driver->amphora queue->amphora endpoint->controller worker. | 18:53 |
| johnsom | They are doing API->driver interface->A10 driver->amphora queue->amphora endpoint->ERROR | 18:53 |
| johnsom | https://github.com/openstack/octavia-lib/blob/master/octavia_lib/api/drivers/provider_base.py#L63 is the driver interface | 18:53 |
| johnsom | If they want to use a queue, they should do API->driver interface->A10 driver->A10 queue->A10 endpoint->A10 code | 18:55 |
| rouk | yeah, for now i just want it working in my lab like it was workign in their lab, then i gotta make a giant list of everything thats wrong | 18:58 |
| rouk | so, goal is working model in lab, then ill try and teach them or make a giant PR | 18:59 |
| rouk | we asked for this, they rushed it and coded it like garbage it seems., | 18:59 |
| *** goldyfruit_ has quit IRC | 19:52 | |
| *** Vorrtex has quit IRC | 19:52 | |
| *** goldyfruit has joined #openstack-lbaas | 19:52 | |
| rouk | johnsom: shouldnt this arg have a default of none or something? seems strange to require it as a mandatory arg, never used to be an arg at all https://github.com/openstack/octavia-lib/blob/master/octavia_lib/api/drivers/provider_base.py#L27 | 19:56 |
| cgoncalves | hope you get a good license deal after all that work ;) | 19:57 |
| johnsom | rouk Good question, rm_work added that, maybe he will have an answer. | 19:58 |
| rouk | cgoncalves: i just want things to not be a mess all the time, and things are a mess all the time. | 19:59 |
| rouk | johnsom: it just seems strange, yet another incompatibility with master with this driver that likes to borrow all the amphora components, but that one does seem weird. | 20:00 |
| rm_work | Lemmie get context | 20:09 |
| rm_work | rouk / johnsom: it was supposed to be used in the multivip stuff, but that patch isn't going to make Train | 20:19 |
| rm_work | so it'll hopefully be in Ussuri | 20:19 |
| rm_work | as to why it's *mandatory*... | 20:20 |
| rm_work | ¯\_(ツ)_/¯ | 20:20 |
| rm_work | all args there are mandatory | 20:20 |
| rm_work | and it's a dict structure so it'd need to be a "None with autoconvert to emptydict" and it made sense just to require the driver to deal with it somehow | 20:21 |
| rm_work | whether that's making it have a default in THEIR signature (which is fine?) or explicitly passing an empty dict to it | 20:21 |
| rm_work | I mean, MAYBE multivip could make Train if I go *hard* on it today and focus 100%, and then convince Michael to do the same to review it, and I don't run into any issues (which I think I will, specifically I never dealt with the random 404 issue we had) | 20:22 |
| *** pcaruana has quit IRC | 20:32 | |
| rm_work | as PTL I would like the power to just vote "YES" and skip zuul and automerge a patch >_> | 20:38 |
| rm_work | since apparently people can do that in other projects... <_< | 20:38 |
| *** trident has quit IRC | 20:39 | |
| rm_work | ala: https://review.opendev.org/#/c/681630/ | 20:39 |
| johnsom | Picking up bad behaviors in the seedy corners of OpenStack? | 20:41 |
| johnsom | Though I have to say, self approving a revert to fix every project is probably a good exception. | 20:43 |
| colin- | is compute_flavor available to be set through the flavor or flavor profile apis? | 20:45 |
| colin- | didn't see it mentioned here https://docs.openstack.org/octavia/stein/admin/flavors.html, but do see it as a property when i use "openstack loadbalancer provider capability list amphora" | 20:45 |
| colin- | (could i supply a pair like --flavor-data '{"compute_flavor": "c4.xlarge"}' | 20:48 |
| johnsom | https://www.irccloud.com/pastebin/y9tSzG8v/ | 20:48 |
| colin- | yeah it was seeing the second property in that output that made me curious | 20:50 |
| *** trident has joined #openstack-lbaas | 20:51 | |
| johnsom | Lol, well, the master docs maybe make it more clear? https://docs.openstack.org/octavia/latest/admin/flavors.html | 20:54 |
| *** ajay33 has quit IRC | 21:16 | |
| colin- | not exactly, the part i'm interested in is summarized as "the amphora provider supports the loadbalancer_topology capability, among many others:" | 21:21 |
| *** ramishra has quit IRC | 21:23 | |
| *** henriqueof1 has quit IRC | 21:31 | |
| *** henriqueof has joined #openstack-lbaas | 21:40 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Generate PDF documentation https://review.opendev.org/667249 | 21:44 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Generate PDF documentation https://review.opendev.org/667249 | 22:02 |
| *** rcernin has joined #openstack-lbaas | 22:19 | |
| *** boden has quit IRC | 22:28 | |
| *** ivve has quit IRC | 22:52 | |
| *** tkajinam has joined #openstack-lbaas | 23:03 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Generate PDF documentation https://review.opendev.org/667249 | 23:23 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Generate PDF documentation https://review.opendev.org/667249 | 23:57 |
| johnsom | Ok, I think that is as good as it is going to get until we figure out why a few chapters are empty | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!