| *** yamamoto has joined #openstack-lbaas | 00:22 | |
| *** yamamoto has quit IRC | 00:47 | |
| *** yamamoto has joined #openstack-lbaas | 00:55 | |
| *** nicolasbock has quit IRC | 01:04 | |
| openstackgerrit | Brian Haley proposed openstack/octavia master: Remove all usage of six library https://review.opendev.org/701290 | 01:10 |
|---|---|---|
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add oslo middleware healthcheck to Octavia API https://review.opendev.org/711127 | 01:12 |
| openstackgerrit | Brian Haley proposed openstack/octavia master: Remove all usage of six library https://review.opendev.org/701290 | 01:32 |
| *** sapd1 has quit IRC | 04:12 | |
| *** sapd1 has joined #openstack-lbaas | 04:12 | |
| *** vishalmanchanda has joined #openstack-lbaas | 04:57 | |
| *** gcheresh has joined #openstack-lbaas | 05:29 | |
| *** yamamoto has quit IRC | 05:31 | |
| *** yamamoto has joined #openstack-lbaas | 05:45 | |
| *** gcheresh has quit IRC | 06:15 | |
| *** ataraday_ has quit IRC | 06:18 | |
| *** gcheresh has joined #openstack-lbaas | 06:41 | |
| *** yamamoto has quit IRC | 06:49 | |
| *** yamamoto has joined #openstack-lbaas | 06:53 | |
| *** gcheresh has quit IRC | 07:14 | |
| cgoncalves | dawzon, alternatively and maybe easier: 1) clone octavia Git repository, 2) git review -d $YOUR_GERRIT_CHANGE_ID_IN_THE_URL | 07:32 |
| cgoncalves | for example, to download https://review.opendev.org/#/c/711376/, run "git review -d 711376" | 07:33 |
| *** gcheresh has joined #openstack-lbaas | 08:03 | |
| *** maciejjozefczyk has joined #openstack-lbaas | 08:05 | |
| *** ataraday_ has joined #openstack-lbaas | 08:05 | |
| *** ccamposr has quit IRC | 08:07 | |
| ataraday_ | rm_work, Please revisit https://review.opendev.org/#/c/709696/ :) | 08:07 |
| *** maciejjozefczyk has quit IRC | 08:07 | |
| *** ccamposr has joined #openstack-lbaas | 08:07 | |
| *** maciejjozefczyk has joined #openstack-lbaas | 08:09 | |
| *** tkajinam has quit IRC | 08:15 | |
| *** tesseract has joined #openstack-lbaas | 08:17 | |
| *** TMM has quit IRC | 08:19 | |
| *** TMM has joined #openstack-lbaas | 08:19 | |
| *** gcheresh has quit IRC | 09:16 | |
| *** gcheresh has joined #openstack-lbaas | 09:24 | |
| *** yamamoto has quit IRC | 09:25 | |
| *** rpittau|afk is now known as rpittau | 09:38 | |
| *** threestrands has quit IRC | 09:56 | |
| *** ramishra has quit IRC | 10:36 | |
| *** ramishra has joined #openstack-lbaas | 10:43 | |
| *** nicolasbock has joined #openstack-lbaas | 11:05 | |
| *** yamamoto has joined #openstack-lbaas | 11:27 | |
| *** rpittau is now known as rpittau|afk | 11:34 | |
| *** rpittau|afk is now known as rpittau|bbl | 11:34 | |
| *** nicolasbock has quit IRC | 11:34 | |
| *** tkajinam has joined #openstack-lbaas | 11:37 | |
| *** nicolasbock has joined #openstack-lbaas | 11:38 | |
| *** tkajinam has quit IRC | 11:39 | |
| *** tkajinam has joined #openstack-lbaas | 11:40 | |
| *** nicolasbock has quit IRC | 11:40 | |
| *** tkajinam has quit IRC | 11:41 | |
| *** tkajinam has joined #openstack-lbaas | 11:41 | |
| *** tkajinam has quit IRC | 11:42 | |
| *** tkajinam has joined #openstack-lbaas | 11:53 | |
| *** gcheresh has quit IRC | 12:06 | |
| *** yamamoto has quit IRC | 12:12 | |
| *** rcernin has quit IRC | 12:26 | |
| *** gcheresh has joined #openstack-lbaas | 12:27 | |
| *** gcheresh has quit IRC | 12:32 | |
| *** nicolasbock has joined #openstack-lbaas | 12:42 | |
| *** rpittau|bbl is now known as rpittau | 12:51 | |
| *** gcheresh has joined #openstack-lbaas | 12:53 | |
| rm_work | I find git review command needs a little more setup sometimes, the checkout commands are more reliable | 13:45 |
| rm_work | But yes, on your main dev system, definitely use `git review -d CR#` | 13:46 |
| rm_work | Ok ataraday_ ! I have a few urgent things for this morning but I'll look ASAP | 13:46 |
| johnsom | cgoncalves Where did you see the error on https://review.opendev.org/#/c/711127/6/octavia/common/config.py | 14:16 |
| johnsom | Ah, maybe when it is disabled..... Yeah, let me test that. | 14:16 |
| johnsom | This is what I was fighting with and brought up at the oslo meeting. How oslo middleware handles the config settings is .... poor at best | 14:17 |
| cgoncalves | I get it even when healthcheck is enabled | 14:17 |
| johnsom | Really? can you past a log traceback? | 14:17 |
| johnsom | It passes for me and in the gates, so curious | 14:18 |
| cgoncalves | http://paste.openstack.org/show/790790/ | 14:19 |
| johnsom | Thank you | 14:19 |
| cgoncalves | octavia.conf: https://paste.centos.org/view/e44ec972 | 14:20 |
| *** yamamoto has joined #openstack-lbaas | 14:24 | |
| johnsom | Yeah, ok, so I may need to go back to my old design for that. Bummer. Easy enough though. It's because you do have it enabled. | 14:25 |
| johnsom | cgoncalves Thanks for the review! | 14:25 |
| *** yamamoto has quit IRC | 14:25 | |
| *** TrevorV has joined #openstack-lbaas | 14:27 | |
| cgoncalves | o/ | 14:28 |
| *** gmann is now known as gmann_afk | 14:48 | |
| *** tkajinam has quit IRC | 14:48 | |
| *** gcheresh has quit IRC | 14:59 | |
| *** gcheresh has joined #openstack-lbaas | 15:13 | |
| rm_work | ataraday_: +W :) | 15:21 |
| rm_work | johnsom / cgoncalves: anything I should look at today at the top of your lists? I'll have time to review a couple of things today | 15:23 |
| rm_work | ah, and speaking of which, i found this patch that could use a review still from johnsom: https://review.opendev.org/#/c/705165 | 15:24 |
| johnsom | https://review.opendev.org/713039 https://review.opendev.org/713016 https://review.opendev.org/713032 | 15:24 |
| johnsom | https://review.opendev.org/705317 grin | 15:24 |
| johnsom | https://review.opendev.org/711127 This one, but I still need to fix it for the issue Carlos found. | 15:24 |
| johnsom | https://review.opendev.org/701290 probably | 15:25 |
| rm_work | wow starting me off with some easy ones, eh | 15:25 |
| johnsom | Though I don't understand by we need a non-utf-8 conversion there | 15:25 |
| johnsom | Yeah, just the ones I know have +2 already | 15:26 |
| johnsom | AH, i have a fix needed on https://review.opendev.org/#/c/713032/ too. But you could do that if you want. | 15:26 |
| johnsom | Stuck in a meeting for a while | 15:27 |
| johnsom | Actually I will fix those | 15:28 |
| *** gcheresh has quit IRC | 15:28 | |
| rm_work | ah i already +W'd that one, you mean adding CONTRIBUTING ? | 15:28 |
| johnsom | Yeah | 15:28 |
| rm_work | not super concerned | 15:28 |
| johnsom | No biggy | 15:28 |
| rm_work | it's passing | 15:28 |
| johnsom | The path mentioned is part of the docs build process, so fine | 15:29 |
| rm_work | yes i commented as much | 15:29 |
| rm_work | i recognize it | 15:29 |
| johnsom | Ah, you commented the same | 15:29 |
| rm_work | if it's passing, I don't think CONTRIBUTING matters | 15:29 |
| rm_work | it's never going to change so | 15:29 |
| johnsom | Yeah, can get added later | 15:29 |
| rm_work | zero concern | 15:30 |
| rm_work | so you updated this healthcheck endpoint to actually do something? | 15:30 |
| johnsom | That goal might be a good time to fix that | 15:30 |
| johnsom | I did | 15:30 |
| johnsom | We can probably add another plugin later that checks the rabbit stuff, but this will at least actually check the API process and DB | 15:31 |
| rm_work | yeah, i would love to also have a RMQ connectivity sanity check there | 15:31 |
| rm_work | that's a big part of the failures I've had <_< | 15:31 |
| johnsom | Before it would *always* be happy if the wsgi even started | 15:31 |
| rm_work | yeah lol | 15:31 |
| rm_work | FYI: https://review.opendev.org/#/c/713296/ | 15:32 |
| rm_work | fixed senlin a bit | 15:32 |
| johnsom | Yeah, on rabbit,we need to figure out *how* to do that check with oslo messaging in the way. | 15:32 |
| rm_work | i don't think we NEED to send a message, I think we can just initiate the connection and it'll give us an immediate failure or not? | 15:32 |
| rm_work | hmm, maybe not | 15:33 |
| johnsom | Yeah, that is the question of *how*. Really I don't want it to go to the backends, just can it get to rabbit from the API | 15:33 |
| rm_work | right | 15:33 |
| johnsom | Once it is *in* rabbit, then that is a different monitor/check IMO | 15:33 |
| rm_work | ugh my task this week is to rebase and address comments on multi-vip, and get testing running on it again | 15:33 |
| rm_work | but i might ditch that | 15:34 |
| johnsom | Not to mention, the backends are random, so not a good test | 15:34 |
| rm_work | if we think we're going to cut that from ussuri goals | 15:34 |
| openstackgerrit | Merged openstack/python-octaviaclient master: Remove the dependency on the "mock" package https://review.opendev.org/713016 | 15:39 |
| johnsom | I would like to see it, but it's a conflict monster | 15:40 |
| rm_work | yes and I really need your help actually | 15:42 |
| rm_work | so I need to wait until we get failover in and you're freed up a bit | 15:42 |
| rm_work | as I'd much rather see that done | 15:42 |
| rm_work | failover fix is a huge deal :D | 15:42 |
| rm_work | i will start looking today | 15:42 |
| rm_work | can you merge https://review.opendev.org/#/c/705165? :D | 15:43 |
| rm_work | oh, wait... that actually conflicts with failover flows, rofl | 15:43 |
| johnsom | Ok, thanks. I still have some polish to do, but it should be good enough for a first pass review. (maybe some strings->constants, stuff like that left) | 15:43 |
| openstackgerrit | Merged openstack/python-octaviaclient master: Align python-octaviaclient to octavia coding style https://review.opendev.org/713032 | 15:44 |
| johnsom | Build me a list of what you want me to look at. Booked this morning, but will try to get things. | 15:44 |
| rm_work | ok the list looks like: https://review.opendev.org/#/c/705165 https://review.opendev.org/#/c/699521 | 15:44 |
| rm_work | followed by those two about 100 times | 15:44 |
| rm_work | https://review.opendev.org/#/q/topic:az-tweaks+(status:open+OR+status:merged) for ease of reference | 15:45 |
| rm_work | huh, don't care so much about the merged ones, thanks gerrit | 15:45 |
| rm_work | yeah ok the six patch looks much better, thanks haleyb | 15:57 |
| rm_work | i'm ok with this now | 15:57 |
| haleyb | rm_work: i think johnsom had a comment on the latin-1 thing | 15:58 |
| rm_work | yes i commented too | 15:58 |
| rm_work | i ALMOST told you to fix it earlier but i looked into it and i get why | 15:58 |
| haleyb | will look... | 15:58 |
| rm_work | it's going to be necessary as long as we're doing our cert parsing the way we do using low-level libs | 15:58 |
| rm_work | they take/return binary data that we need to deal with | 15:59 |
| rm_work | it's not str/unicode, which is the py2-3 thing | 15:59 |
| rm_work | it's binary/string | 15:59 |
| rm_work | which is still a thing even in py3 | 15:59 |
| johnsom | Yeah, if it is exceptions for certain tools, ok. But even when generating certs I would question if non-utf-8 is really the right answer | 15:59 |
| rm_work | oh, i mean if you just want to switch to utf8 sure | 16:00 |
| rm_work | i tested that and even the hexdigest stuff is the same with md5 | 16:00 |
| rm_work | whether you encode latin1 or utf8 | 16:00 |
| johnsom | I think the Octavia way from day one was UTF-8 first | 16:00 |
| rm_work | it's PROBABLY fine to switch | 16:00 |
| rm_work | right it was just an abundance of caution, because the CURRENT code uses latin1 | 16:00 |
| rm_work | because we used six to do it and six used that behind the scenes | 16:00 |
| rm_work | so the no-change way is to keep that | 16:00 |
| rm_work | but i don't THINK we ever used anything that would cause it to make a difference, so it'd probably be safe to change that to utf8 | 16:01 |
| *** gcheresh has joined #openstack-lbaas | 16:03 | |
| johnsom | Yeah, I know we got burnt with a multi-byte string early, I think for a "name" field, so we agreed to focus on UTF-8 everywhere. (also to hopefully make the py3 change easier). So, probably a good time to double check "is this really 8bit" again. | 16:09 |
| *** gcheresh has quit IRC | 16:12 | |
| *** gcheresh has joined #openstack-lbaas | 16:26 | |
| *** servagem has joined #openstack-lbaas | 16:30 | |
| rm_work | want me to fix https://review.opendev.org/#/c/711127/ ? | 16:53 |
| rm_work | whatever, just gonna do it ^^ FYI | 16:53 |
| *** rpittau is now known as rpittau|afk | 17:10 | |
| *** gcheresh has quit IRC | 17:20 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add oslo middleware healthcheck to Octavia API https://review.opendev.org/711127 | 17:23 |
| johnsom | rm_work ummm, I fixed it | 17:24 |
| johnsom | Didn't see your message | 17:24 |
| rm_work | yeah no worries, i was in-progress but got distracted by internal explosion | 17:24 |
| johnsom | Figured out I didn't fix the app conf when I moved the test stuff around. I also found another bug in the functional tests.... | 17:24 |
| johnsom | Missed a config override, so my local config got pulled in for the functional test. Fixed it. | 17:26 |
| johnsom | Ah, I can fix that workaround too. Just a sec | 17:30 |
| openstackgerrit | Merged openstack/octavia master: [Amphorav2] Fix noop driver case https://review.opendev.org/709696 | 17:36 |
| openstackgerrit | Merged openstack/octavia-lib master: Remove the dependency on the "mock" package https://review.opendev.org/713039 | 17:36 |
| *** tesseract has quit IRC | 17:37 | |
| johnsom | Ok, running the tests and I will push a version that fixes that pecan import issue. | 17:52 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add oslo middleware healthcheck to Octavia API https://review.opendev.org/711127 | 17:54 |
| johnsom | Mixing import and from x import can lead to this partial module loading bug in python | 17:55 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add oslo middleware healthcheck to Octavia API https://review.opendev.org/711127 | 18:07 |
| johnsom | Ha, self review realized I wasn't consistent in the import naming. | 18:07 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add oslo middleware healthcheck to Octavia API https://review.opendev.org/711127 | 18:08 |
| rm_work | huh, have a link on that bug? sounds interesting | 18:11 |
| johnsom | I tracked it down once. it has to do with how the module loader "caches" modules names it has "loaded". So if a partial, i.e. "from x import y" is first, x.y is imported, module x is cached as loaded, but the rest of module x is not actually loaded. so a later file with "import x" will cause the module loader to say "done" and move on. So references to x.z in that file will fail. | 18:17 |
| johnsom | If you want an example, checkout https://review.opendev.org/#/c/711127/7 (i.e. version 7), then delete line 22 the pecan rest import. | 18:18 |
| johnsom | restart the api and it will blow chunks with rest not found | 18:19 |
| johnsom | Note that rest isn't used in that file. The error will be rest not found in other files | 18:19 |
| rm_work | I was right, that is interesting :D | 18:35 |
| johnsom | Burned many brain cells when I first hit it. | 18:35 |
| johnsom | I use it as an interview question now. If they know about it, they have done significant work in python.... | 18:36 |
| rm_work | lol I often ask the mutable default args one | 18:39 |
| rm_work | But at the same time, I feel like it's a bit mean to ask questions where it's entirely possible for someone to just never run into something and still be quite skilled, and then judge them on it | 18:40 |
| rm_work | But I think that one is likely enough to come up that it's ok | 18:40 |
| johnsom | Oh yeah, that is just one of many clues you get during the interview | 19:06 |
| *** vishalmanchanda has quit IRC | 19:07 | |
| *** gmann_afk is now known as gmann | 19:25 | |
| *** openstackgerrit has quit IRC | 19:33 | |
| *** maciejjozefczyk has quit IRC | 19:39 | |
| TrevorV | Hey guys, been a while (again), but quick-ish question: Why is "acpid" installed on the amphora-agent? | 19:51 |
| TrevorV | I can't seem to find it specifically used anywhere in the code, but I'm also probably looking for it incorrectly | 19:52 |
| johnsom | It is not used by Octavia directly, it comes in the base cloud images (linux). The hypervisor will use it for controlling some actions on the VM. For example, if you stop the VM via kvm or libvirt, that signal can be (depending on how you call it) passed from the hypervisor, kvm, down to the guest linux running the amphora code. | 19:58 |
| TrevorV | So its a hard requirement to be used on *any* architecture amphorae, is that right? | 19:59 |
| johnsom | No, again the amphorae don't care, it's the OS in the VM that cares. It is also typically only on x86 architecture OSes. | 20:00 |
| johnsom | Though I think there are exceptions. | 20:00 |
| johnsom | https://en.wikipedia.org/wiki/Advanced_Configuration_and_Power_Interface | 20:00 |
| TrevorV | Oh oh I gotcha, sorry, I misread that as "needed", my fault. Alright. I might have my first octavia patch in a while coming soon. | 20:01 |
| johnsom | If the OS the amphora is running doesn't have it, we will not care nor notice it is gone | 20:01 |
| TrevorV | I was reading through that already, just didn't find anything I was seeing in the octavia codebase, that's all :) | 20:01 |
| johnsom | Right, it's not an Octavia thing at all | 20:02 |
| TrevorV | Cool, thanks a bunch! | 20:02 |
| johnsom | Purely an OS thing | 20:02 |
| dawzon | How do you install octavia-lib from source? I couldn't find anything in the docs | 20:05 |
| johnsom | If you have it cloned, cd into the directory, then "pip install ." (there is a period there) | 20:05 |
| *** TrevorV has quit IRC | 20:25 | |
| *** openstackgerrit has joined #openstack-lbaas | 20:39 | |
| openstackgerrit | Michael Johnson proposed openstack/python-octaviaclient master: Fix doc8 check https://review.opendev.org/713531 | 20:39 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Remove the barbican "Grant access" from cookbook https://review.opendev.org/713533 | 20:48 |
| openstackgerrit | Michael Johnson proposed openstack/octavia stable/train: Remove the barbican "Grant access" from cookbook https://review.opendev.org/713535 | 20:53 |
| openstackgerrit | Michael Johnson proposed openstack/octavia stable/train: Remove the barbican "Grant access" from cookbook https://review.opendev.org/713535 | 20:55 |
| openstackgerrit | Michael Johnson proposed openstack/octavia stable/stein: Remove the barbican "Grant access" from cookbook https://review.opendev.org/713536 | 20:56 |
| openstackgerrit | Michael Johnson proposed openstack/octavia stable/rocky: Remove the barbican "Grant access" from cookbook https://review.opendev.org/713537 | 20:56 |
| johnsom | If you haven't guessed, I'm working with our downstream docs guy a bit today.... lol | 20:57 |
| johnsom | rm_work FYI: https://bugs.launchpad.net/ubuntu/+source/python-barbicanclient/+bug/1867676 | 21:06 |
| openstack | Launchpad bug 1867676 in python-barbicanclient (Ubuntu Bionic) "Fetching by secret container doesn't raises 404 exception" [Undecided,Confirmed] | 21:06 |
| johnsom | The issue with the fallback to the legacy barbican container was a bug in barbican client. | 21:07 |
| johnsom | Or two I guess | 21:07 |
| *** rcernin has joined #openstack-lbaas | 21:30 | |
| *** rcernin has quit IRC | 21:37 | |
| *** rcernin has joined #openstack-lbaas | 21:38 | |
| sorrison | Are there any examples/docs on how to use the octaviaclient (not CLI reference) can't seem to find anything and it looks like the client is set out differently to other openstack clients | 21:47 |
| johnsom | There is no "octaviaclient" only python-openstackclient and the Octavia plugin. | 21:56 |
| johnsom | So, when you install python-openstackclient, then install python-octaviaclient. You will then have "openstack loadbalancer <...>" available that aligns to all of the other openstack services. | 21:57 |
| sorrison | I am building a python application | 21:58 |
| sorrison | what is python-octaviaclient then? | 21:58 |
| johnsom | The separate clients were all going deprecated, so we never created an "octaviaclient", just the OpenStack client plugin. | 21:59 |
| johnsom | Oh, so you are looking for an SDK? | 21:59 |
| sorrison | https://github.com/openstack/python-octaviaclient/blob/master/octaviaclient/api/v2/octavia.py | 21:59 |
| sorrison | I don't know exactly, I just want a client library for octavia | 22:00 |
| sorrison | usually that means python-<project>client | 22:00 |
| johnsom | The python SDK for the openstack services is openstacksdk | 22:00 |
| johnsom | https://opendev.org/openstack/openstacksdk | 22:00 |
| johnsom | https://docs.openstack.org/openstacksdk/latest/ | 22:00 |
| sorrison | yeah ok, haven't used that yet but will give it a go, all our code uses the client libs as it was pre SDK when we started. | 22:01 |
| johnsom | Yeah, all of those proprietary clients are basically deprecated. The /api/v2/octavia.py is there to support the OSC plugin until we switch it over to openstacksdk | 22:01 |
| sorrison | ok ta | 22:01 |
| johnsom | Yeah, we got caught up in the transition, so.... | 22:01 |
| sorrison | I love the openstack client space...... | 22:02 |
| sorrison | yeah fun times :-) | 22:02 |
| johnsom | At least the openstackclient is clean and unified. The commands for the most part work the same, etc. | 22:02 |
| sorrison | esp when glance and nova says they only support python-glanceclient and not openstacksdk/osc | 22:02 |
| johnsom | Yeah, well, you probably saw the e-mail list on that | 22:03 |
| johnsom | Careful or you will get rm_work fired up on that topic too.... grin | 22:03 |
| sorrison | missed that but I can assume what it was about | 22:03 |
| johnsom | Yeah. | 22:03 |
| johnsom | We ditched the legacy glanceclient a long time ago and have not missed it. | 22:04 |
| sorrison | about time I learnt the SDK anyway, I was going to put in a patch to it for supporting AZ attribute in LBs but someone beat me to it :-) | 22:04 |
| sorrison | Aiming to update the horizon plugin to support AZs too, unless you or rm_work working on that already? | 22:05 |
| johnsom | rm_work may have had a conversation about that with Jacky. I can't remember. | 22:06 |
| dawzon | Does it make sense to set the TLS cipher field even for listeners that aren't TERMINATED_HTTPS? Or would it be better to fill in the field at the time when a listener is switched to https-terminated mode? | 22:06 |
| johnsom | He is our resident dashboard expert | 22:06 |
| sorrison | ok I'll hold off starting something until I hear from them | 22:07 |
| johnsom | dawzon If the listener is created with non-terminated-tls I expect the TLS cipher field to be None | 22:07 |
| johnsom | dawzon Also, you can only specify the protocol at listener create time, it can't be changed after it is created via the API | 22:08 |
| dawzon | Oh. That makes my job a bit easier | 22:09 |
| johnsom | Yep | 22:09 |
| johnsom | The ciphers can be changed on a terminated HTTPS listener though | 22:10 |
| johnsom | I.e. they can be updated once the listener is already created with HTTPS | 22:10 |
| johnsom | sorrison FYI, the dashboard uses openstacksdk already if you want a reference. | 22:11 |
| johnsom | https://github.com/openstack/octavia-dashboard/blob/master/octavia_dashboard/api/rest/lbaasv2.py | 22:12 |
| sorrison | all good, I'm on my way | 22:12 |
| sorrison | johnsom: Actually a little stuck, can't seem to find a way to get/set LB quotas | 22:19 |
| sorrison | doesn't seem to be a `get_loadbalancer_quotas` like rest of projects | 22:19 |
| sorrison | maybe this isn't supported with sdk? | 22:19 |
| sorrison | I can see https://github.com/openstack/openstacksdk/blob/master/openstack/load_balancer/v2/quota.py | 22:20 |
| johnsom | Yeah, was just going to point to that | 22:20 |
| sorrison | so it's there somewhere but I'm struggling to find how | 22:20 |
| johnsom | Though I don't see it in the docs, so maybe the docs rendering is broken | 22:20 |
| sorrison | eg for neutron quota there is https://github.com/openstack/openstacksdk/blob/master/openstack/cloud/_network.py#L637 | 22:21 |
| johnsom | Hmm, well, here is the patch: https://review.opendev.org/#/c/590193/ | 22:22 |
| sorrison | I see get_network_quota is returning LB quotas so maybe this hasn't been split out to it's own module yet | 22:23 |
| johnsom | Yeah, that is a mistake/bug on the neutron side | 22:23 |
| johnsom | Someone added a bunch of strange quotas to neutron (like lbaasv1 stuff) and they merged it.... | 22:24 |
| johnsom | I think they are rolling that back now, don't know. | 22:24 |
| johnsom | I think Brian had a patch up to remove some of that bogus quota stuff from the neutron side | 22:24 |
| johnsom | Anyway, back on Octavia, I see methods on the proxy in the patch: https://review.opendev.org/#/c/590193/20/openstack/load_balancer/v2/_proxy.py | 22:25 |
| johnsom | Those aren't working for you? | 22:25 |
| sorrison | ahh I figured it out now | 22:26 |
| johnsom | Ah, yeah, the docs links didn't get updated for quota.... I'll push a patch to fix that | 22:26 |
| sorrison | connection.load_balancer.get_quota() sorry it's slightly confusing as some stuff is connection.get_network_quotas etc. | 22:26 |
| johnsom | Yeah, just FYI, neutron/networking never had any features for Octavia there, so if you see something, it is a mirage, ignore it | 22:27 |
| sorrison | lol | 22:27 |
| sorrison | johnsom: so did you say there are docs for LB quotas somewhere but just not being rendered? The generated docs look wrong as they say they accept 2 params for the update_quota() method but it only accepts one and I'm struggling to figure out how to set a quota | 22:51 |
| johnsom | Yeah, I just put a patch together and was about to look at how it renders. The Octavia quotas are just missing from the SDK docs at the moment. | 22:52 |
| johnsom | Give me 10 minutes | 22:52 |
| sorrison | ok thanks heaps! | 22:52 |
| johnsom | https://www.irccloud.com/pastebin/nJ3esRLo/ | 22:55 |
| johnsom | That can get you by until zuul does it's thing on: https://review.opendev.org/713551 | 22:56 |
| sorrison | It's not clear how to update a quota | 22:57 |
| sorrison | I tried: client.load_balancer.update_quota('project-id', {'loadbalancer': 10}) | 22:57 |
| sorrison | gives: update_quota() takes 2 positional arguments but 3 were given | 22:58 |
| johnsom | blah, missed that the "resource" is missing too | 22:58 |
| *** tkajinam has joined #openstack-lbaas | 22:58 | |
| johnsom | sorrison You need to create the "quota" resource, then call update using the resource. | 23:01 |
| johnsom | https://opendev.org/openstack/openstacksdk/src/branch/master/openstack/load_balancer/v2/quota.py | 23:01 |
| sorrison | ahhh ok | 23:02 |
| sorrison | now to figure out how to instantiate one of those | 23:04 |
| johnsom | Yeah, I am a bit rusty as well. It was on my recent "I should go look at that again" list. | 23:07 |
| sorrison | and looks like it will take in any argument you give it, haven't been able to guess it yet. Error from octavia is just a 400, Tried load_balancer, loadbalancer, load_balancers, load_balancer | 23:08 |
| sorrison | q = quota.Quota(project_id='0bdf024c921848c4b74d9e69af9edf08', load_balancers=10) | 23:08 |
| sorrison | client.load_balancer.update_quota(q) | 23:08 |
| sorrison | BadRequestException: 400 | 23:08 |
| sorrison | this stuff is so hard | 23:09 |
| sorrison | It seems load_balancers is the right attribute for the Quota resource but octavia still throwing a 400 | 23:10 |
| johnsom | So, if I read the doc string, you get a qouta resource via get_quota, then use that to call update_quota(<resource>, load_balancers=10) | 23:11 |
| johnsom | get_quota(<project_id>) | 23:12 |
| sorrison | success!!! | 23:13 |
| johnsom | Cool. The docs update is merging now | 23:13 |
| sorrison | kinda strange it requires 2 api calls but hey it works | 23:13 |
| sorrison | thanks for you help | 23:14 |
| johnsom | NP | 23:14 |
| sorrison | The other confusing thing is that the string representations of the quota resources say openstack.load_balancer.v2.quota.Quota(id=pt-50, load_balancer=10, listener=-1, member=-1, pool=-1, health_monitor=-1, | 23:21 |
| sorrison | but the attributes are actually load_balancers, listeners etc. | 23:22 |
| johnsom | Yeah, I would call that a bug in that patch, though I know some of the SDK stuff has strange pluralization, so maybe that is an SDK thing??? | 23:22 |
| johnsom | Feel free to file an SDK bug for that | 23:23 |
| *** ccamposr has quit IRC | 23:27 | |
| *** ccamposr has joined #openstack-lbaas | 23:28 | |
| *** ccamposr has quit IRC | 23:52 | |
| *** ccamposr has joined #openstack-lbaas | 23:53 | |
| *** ccamposr has quit IRC | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!