Thursday, 2020-03-19

« Wednesday, 2020-03-18 Index Friday, 2020-03-20 »
*** ramishra has quit IRC00:04
*** nicolasbock has quit IRC00:28
*** vishalmanchanda has joined #openstack-lbaas01:33
*** ianychoi_ has joined #openstack-lbaas01:38
*** ianychoi has quit IRC01:40
*** yamamoto has joined #openstack-lbaas01:43
*** sapd1 has joined #openstack-lbaas01:50
*** sapd1_x has joined #openstack-lbaas01:52
*** servagem has quit IRC01:58
*** yamamoto has quit IRC02:16
*** yamamoto has joined #openstack-lbaas02:17
*** psachin has joined #openstack-lbaas02:31
*** yamamoto has quit IRC03:00
*** yamamoto has joined #openstack-lbaas03:01
*** ramishra has joined #openstack-lbaas03:29
rm_worksorrison: it was on my list, but very far down it. If you can, that'd be great. Possibly also tho, still need to get the AZ tempest testing in better shape and merged... Need to actually have a different network set up, so we can prove the fixes I have up are necessary (rather than just a second AZ on all the same networks)04:06
sorrisonOK cool, we can take care of the dashboard. We're just about to hire an ex horizon core :-)04:08
sorrisonand yeah forgot about tempest changes, will try tackle them next week04:09
*** yamamoto has quit IRC04:17
*** yamamoto has joined #openstack-lbaas04:28
dawzonI'm having some trouble updating a listener over REST with curl.  Getting 405 method not allowed back, is there something I'm missing?  Here's my command: curl -X PUT -H "Content-Type: application/json" -H "X-Auth-Token: <token>" -d {"listener": {"tls_ciphers": "SHA384:ECDHE-RSA-AES128-GCM-SHA256"}} localhost/v2/lbaas/listeners/17a23f66-282b-4fe1-aeaa-0a779e2a364404:38
*** ramishra has quit IRC04:54
*** ramishra has joined #openstack-lbaas05:01
johnsomDid your token expire?05:24
*** gcheresh has joined #openstack-lbaas05:51
*** gcheresh has quit IRC06:00
*** gcheresh has joined #openstack-lbaas06:19
*** gcheresh has quit IRC06:25
*** gcheresh has joined #openstack-lbaas06:46
*** gcheresh has quit IRC07:31
*** gcheresh has joined #openstack-lbaas07:33
*** tesseract has joined #openstack-lbaas07:51
*** maciejjozefczyk has joined #openstack-lbaas07:58
*** threestrands has quit IRC08:01
*** dawzon has quit IRC08:07
*** lxkong has quit IRC08:08
*** kevinz has quit IRC08:08
*** luketollefson has quit IRC08:09
*** guilhermesp has quit IRC08:09
*** fyx has quit IRC08:09
*** luketollefson has joined #openstack-lbaas08:09
*** dawzon has joined #openstack-lbaas08:09
*** lxkong has joined #openstack-lbaas08:10
*** mnaser has quit IRC08:10
*** kevinz has joined #openstack-lbaas08:11
*** fyx has joined #openstack-lbaas08:11
*** mnaser has joined #openstack-lbaas08:11
*** rpittau|afk is now known as rpittau08:11
*** guilhermesp has joined #openstack-lbaas08:11
*** tkajinam has quit IRC08:12
*** yamamoto has quit IRC08:16
*** yamamoto has joined #openstack-lbaas08:30
*** psachin has quit IRC09:01
*** psachin has joined #openstack-lbaas09:02
*** yamamoto has quit IRC09:04
*** dawzon has quit IRC09:08
*** dawzon has joined #openstack-lbaas09:09
*** vishalmanchanda has quit IRC10:03
*** vishalmanchanda has joined #openstack-lbaas10:22
*** yamamoto has joined #openstack-lbaas10:32
*** tesseract has quit IRC10:36
openstackgerritMaciej Józefczyk proposed openstack/octavia-tempest-plugin master: Do not use requests.session in _wait_for_lb_functional()  https://review.opendev.org/71382610:38
openstackgerritMaciej Józefczyk proposed openstack/octavia-tempest-plugin master: Send HTTP request while testing L4 TCP Load Balancer  https://review.opendev.org/71382710:38
openstackgerritMaciej Józefczyk proposed openstack/octavia-tempest-plugin master: Do not test reuse connections for UDP  https://review.opendev.org/71382810:38
*** tesseract has joined #openstack-lbaas10:42
*** tesseract has quit IRC10:46
*** tesseract has joined #openstack-lbaas10:48
*** yamamoto has quit IRC10:52
*** ivve has joined #openstack-lbaas10:55
openstackgerritMaciej Józefczyk proposed openstack/octavia-tempest-plugin master: Do not test reuse connections for UDP  https://review.opendev.org/71382810:56
*** tesseract-RH has joined #openstack-lbaas11:00
*** tesseract-RH has left #openstack-lbaas11:02
*** tesseract has quit IRC11:03
*** sapd1_x has quit IRC11:04
*** nicolasbock has joined #openstack-lbaas11:05
*** sapd1 has quit IRC11:05
*** pck has joined #openstack-lbaas11:16
*** pck is now known as pckizer11:16
*** rpittau is now known as rpittau|bbl11:30
*** ccamposr__ has joined #openstack-lbaas11:31
*** ccamposr has quit IRC11:33
*** yamamoto has joined #openstack-lbaas11:56
*** gcheresh has quit IRC12:18
*** ccamposr__ has quit IRC12:26
*** ccamposr__ has joined #openstack-lbaas12:26
*** gcheresh has joined #openstack-lbaas12:33
*** yamamoto has quit IRC12:47
*** rpittau|bbl is now known as rpittau12:59
*** gcheresh has quit IRC13:16
*** nicolasbock has quit IRC13:18
*** yamamoto has joined #openstack-lbaas13:19
*** nicolasbock has joined #openstack-lbaas13:20
*** yamamoto has quit IRC13:20
*** psachin has quit IRC13:21
openstackgerritCarlos Goncalves proposed openstack/octavia-tempest-plugin master: Fix skip check in active-standby test  https://review.opendev.org/71387413:25
openstackgerritAdam Harwell proposed openstack/octavia-tempest-plugin master: Remove old api_v1_enabled cruft from job defs  https://review.opendev.org/71387913:35
rm_worknot sure if someone had already done that, if so then whatev, just saw it and was faster to do it than look lol13:35
cgoncalvesmay still be applicable to stable branches where v1 was supported13:37
rm_workahh hmm, let me check how those run13:40
rm_workyeah you may be right, we may have to wait one more release13:41
rm_workI think rocky was the last to contain v1 at all?13:41
cgoncalvesdeprecated in rocky, removed in train13:43
cgoncalveshttps://docs.openstack.org/releasenotes/octavia/rocky.html#deprecation-notes13:43
cgoncalveshttps://docs.openstack.org/releasenotes/octavia/train.html#deprecation-notes13:43
rm_workwait yeah seriously gah, just noticed it existed in stein13:43
rm_workso yeah, gotta wait for stein to fall off too, ffff13:43
rm_workoh wel13:44
cgoncalvesor move the setting to the -$branch jobs13:44
rm_workah, time for today's plug for https://review.opendev.org/#/q/topic:az-tweaks+status:open13:44
rm_workah yeah could do that13:44
*** yamamoto has joined #openstack-lbaas13:55
openstackgerritAdam Harwell proposed openstack/octavia-tempest-plugin master: Remove old api_v1_enabled cruft from job defs  https://review.opendev.org/71387913:56
rm_workfuck, rocky gates are broken for us because of MarkupSafe14:16
rm_work>_<14:16
rm_work2020-03-19 13:55:05.431232 | controller | 2020-03-19 13:55:05.431 |   Downloading MarkupSafe-1.0.tar.gz (14 kB)14:16
rm_work2020-03-19 13:55:05.680071 | controller | 2020-03-19 13:55:05.679 |     ERROR: Command errored out with exit status 1:14:16
rm_work2020-03-19 13:55:05.680140 | controller | 2020-03-19 13:55:05.679 |      command: /opt/amphora-agent-venv/bin/python -c 'import sys, setuptools, tokenize; sys.argv[0] = '"'"'/tmp/pip-install-feio8mx5/MarkupSafe/setup.py'"'"'; __file__='"'"'/tmp/pip-install-feio8mx5/MarkupSafe/setup.py'"'"';f=getattr(tokenize, '"'"'open'"'"', open)(__file__);code=f.read().replace('"'"'\r\n'"'"', '"'"'\n'"'"');f.close();exec(compile(code,14:16
rm_work__file__,14:16
rm_work '"'"'exec'"'"'))' egg_info --egg-base /tmp/pip-install-feio8mx5/MarkupSafe/pip-egg-info14:16
rm_workbeen dealing with that elsewhere for like a week internally14:17
rm_workdidn't realize it would hit us too T_T14:17
rm_workI honestly don't know how to fix this besides updating to MarkupSafe==1.1, which is capped at 1.0 in requirements upper-constraints in rocky14:17
rm_workdownstream we just patch it and ignore the u-c... but we can't do that in the gates... so... ???14:18
rm_workwill ask infra14:18
*** yamamoto has quit IRC14:24
openstackgerritMerged openstack/octavia master: Remove all usage of six library  https://review.opendev.org/70129014:27
openstackgerritMerged openstack/octavia master: Remove the barbican "Grant access" from cookbook  https://review.opendev.org/71353314:27
*** TrevorV has joined #openstack-lbaas14:30
rm_worksee https://zuul.opendev.org/t/openstack/build/b75e42523e6c41e5a6e97c722fbb0ecd for an example of what's broken14:31
rm_workthere's only two options AFAIK: upgrade to MarkupSafe>=1.1, or somehow pin setuptools to an older version... I don't honestly know if it is possible to do either...14:32
johnsomCan't we pin in our requirements.txt file?14:34
johnsomI would ask in requirements myself14:34
rm_worki did previously and posted a patch14:37
rm_workbut discussion was not very deep14:37
rm_workwe can't pin setuptools easily as it's pre-installed in the virtualenv14:38
rm_workit's not something we pip-install14:38
rm_workthat's why it suddenly broke as far back as rocky14:38
johnsomAh, yeah, that is a problem14:38
rm_worki'm honestly a little stumped here, without resorting to some EVIL trickery14:38
rm_worklike bumping the version and just... dropping the requirements job14:39
johnsomWhat did the requirements folks say?14:40
rm_work"no"14:40
rm_worklemmie see if i can find the exact scrollback14:40
rm_workoh huh14:41
rm_workmaybe they changed their minds14:41
* rm_work looks14:41
rm_worksee reqs channel14:43
*** gcheresh has joined #openstack-lbaas14:48
rm_workjohnsom: I may propose dropping support for queens/rocky going forward as the CI is essentially no longer fixable15:09
rm_workcgoncalves: ^^15:09
*** gcheresh has quit IRC15:20
*** gcheresh has joined #openstack-lbaas15:28
*** gcheresh has quit IRC15:35
cgoncalvesrm_work, you mean dropping the queens and rocky jobs in octavia-tempest-plugin CI?15:37
*** Trevor_V has joined #openstack-lbaas15:37
rm_workyes, and declaring queens/rocky unmaintained15:37
rm_workI am making a patch to test one alternative option tho15:37
johnsomPersonally I am itching to drop queens. It's still on xenial, etc.... It is just getting vary dated at this point.15:39
johnsomI think babysitting FIVE releases is..... a bit much15:39
rm_workyes15:39
rm_workwe can apparently just... decide to drop them whenever15:39
rm_workthere's no requirement to continue supporting them15:40
rm_workand it especially makes sense in case of unfixable CI, like... this15:40
johnsomThat said, I am still on the hook for queens downstream, but that is a different issue/different hat to wear15:40
*** TrevorV has quit IRC15:40
*** vishalmanchanda has quit IRC15:43
johnsomYeah, so how do we declare the "Unmaintained" or EOL status for that? This was a question I raised when they made these new categories, but got no clear answer.15:43
johnsomDo we just drop the gate jobs and *ignore* them?15:43
johnsomI guess EOL gets a tag15:43
cgoncalvesnot sure I fully follow. the changes would only impact the octavia-tempest-plugin CI, yes?15:44
cgoncalvesif so, I agree with stop gating on queens and rocky jobs15:44
rm_workcgoncalves: err, also the gate jobs would be unrunable on the stable branches15:45
rm_workso no changes could be merged anyway15:45
rm_worksorry, "would be" -> "are already"15:45
johnsomI think we are discussing a bigger change, where patches to queens would not trigger any job any longer.15:45
cgoncalvesyeah but if someone wants to get something proposed and merged can still do15:45
rm_workno, they can't, because if you propose a change to rocky now it won't merge15:45
cgoncalvesif CI is broken, there is no guarantees from the octavia project team we would fix them15:45
rm_workerr, won't pass gates15:45
rm_workso how would we merge anything?15:45
rm_worknot going to merge changes especially to a stable release without any testing15:46
johnsomRight, the gate jobs would block everything, so it is the same as EOL15:46
rm_workah damn, looks like requirements decided it's fine to merge the bump15:47
rm_workerr i mean, ah, good15:47
johnsomlol15:47
cgoncalvesmerging queens/rocky backport patches without any CI isn't that good either. I wish we could override Zuul's verification vote in such cases15:47
johnsomWell, I still propose we EOL queens15:47
rm_workyeah15:47
rm_workI second that15:47
johnsomcgoncalves we *could* but, does that make sense or bring value?15:48
*** spatel has joined #openstack-lbaas15:48
rm_workhttps://review.opendev.org/#/q/status:open+project:openstack/octavia+branch:stable/queens15:48
rm_workso these are the only things open on queens15:49
rm_workand I don't see us merging any of them15:49
johnsomWell, one less now, abandoned the fix for queens that I could not get working15:49
rm_workI'm going to make an ML thread for ya'll to vote on? or can we just decide it here15:49
rm_workwe have core quorum?15:49
cgoncalvesjohnsom, job results could be first checked and/or patches tested manually and if looks good we override verification15:49
johnsomWe do, but I think a ML thread is the best answer for this. It impacts others beyond the core team15:50
rm_workk15:50
cgoncalveshaving just said this, I don't want to say I'd manually verify every/any queens patch15:50
johnsomcgoncalves With all the spare time the core team has....15:50
johnsomRight15:50
johnsomlol15:50
cgoncalvesright15:50
rm_workyeah I am just *not* going to do that, so15:50
johnsom<see above nag from rm_work about patch reviews>15:50
*** spatel has quit IRC15:51
* johnsom thanks rm_work for reminding us15:51
rm_workYeah, I will continue to helpfully remind you every day until those merge :D15:51
johnsomAs you should15:51
rm_workjust sent a ML email15:55
rm_workshould show up shortly15:56
cgoncalvesIIUC, many of the issues we see in queens and rocky are due to python3 and other flaky stuff where typically we are the first ones to notice (canary!) because blocks us primarily in octavia-tempest-plugin but other projects should also be impacted by. I would say let's start with removing queens and rocky jobs in tempest plugin first15:56
cgoncalveswhenever someone sees a problem with a queens/rocky backport, it might be that we just need to wait (and pray) a bit for someone to fix tempest/dib/devstack/whatever15:57
cgoncalvesif we see the issue persists for an unreasonable time, we could consider disabling queens and rocky CI15:58
cgoncalvesmy 2 cents15:58
rm_workwell, i am not in favor of dropping testing on tempest repo and still having those tests run on the stable branches15:59
rm_workthat's just asking for trouble15:59
johnsom+116:00
johnsomI don't like "half-EOL" things. I just leads to confusion and trouble IMO.16:01
*** ccamposr has joined #openstack-lbaas16:01
johnsomI would rather lean towards we do or we don't.16:01
cgoncalvescode changes in octavia-tempest-plugin does not impact queens/rocky octavia code. it might uncover issues but not negatively impact16:01
cgoncalvesI'm just worried not having CI coverage would lead people to merge whatever without any checking16:02
johnsomBut code changes in tempest-plugin impact the ability to test queens/rocky at all.16:02
cgoncalvesor cores being blamed for holding backport patches16:02
*** ccamposr__ has quit IRC16:03
johnsomYeah, so, if we declare EOL, we don't have that problem.16:03
cgoncalvesif EOL we rather just stop merging any code in those branches16:04
johnsomI mean really, in a way we are already through the 6 month "gates don't work" window for queens. That interface issue has not been fixed in that time window16:04
cgoncalvesdoesn't that go against the branches in extended maintenance (queens, rocky)?16:04
johnsomTo quote the EOL description: "After a project/branch exceeds the time allocation as Unmaintained, or a team decides to explicitly end support for a branch, it will become End of Life."16:04
cgoncalvesok. I was looking for that now. thanks16:05
cgoncalveshttps://docs.openstack.org/project-team-guide/stable-branches.html#end-of-life16:05
johnsomhttps://docs.openstack.org/project-team-guide/stable-branches.html#end-of-life16:05
cgoncalves"and the branch deleted."16:06
johnsomI just feel we give people a false sense of "support" when the reality is that it probably should be EOL.16:06
johnsomYep16:06
cgoncalvesok. upstream hat on, I am supportive of this16:06
johnsomI get that downstream it makes things... trickier/harder. But it's going to happen at some point anyway.16:07
cgoncalvesagreed. I was also considering the pros/cons upstream16:09
*** maciejjozefczyk has quit IRC16:10
cgoncalvesrm_work, are you expecting an email reply to the email you sent to -discuss?16:10
rm_workyes16:15
rm_work+1's from cores please to show quorum16:15
rm_workand I'll wait a week to see if we get any other feedback i guess16:15
rm_workI suppose I could mandate but trying to be a nice PTL :D16:16
johnsomrm_work FYI, there was a discussion of the state of releases in the release team meeting just now. They were wondering why the Octavia ocata jobs were not working. lol. I think we really need to move forward with a strong EOL statement and getting the branches dropped.17:00
*** gcheresh has joined #openstack-lbaas17:19
*** gcheresh has quit IRC17:58
rm_workROFL wat17:59
rm_worki thought we did EOL everything prior to queens17:59
*** rpittau is now known as rpittau|afk18:06
openstackgerritMichael Johnson proposed openstack/octavia master: Add oslo middleware healthcheck to Octavia API  https://review.opendev.org/71112718:24
johnsomMe too18:24
johnsomRebased, but had conflicts, so we get to vote again....18:24
dawzonjohnsom Still getting the 405 it with a brand new token (taken from the "id" field from openstack token issue, right?).  The command I posted yesterday was missing single quotes around the JSON but that wasn't the problem.  Is the REST API hosted on a different port?18:41
johnsomdawzon Use "openstack endpoint list" to see the URL to use18:42
dawzonAh, that was it.  The URL that gave for octavia is different than the one in the curl example file.  Is that just because I'm running DevStack instead of a regular deployment?18:45
johnsomYeah, the docs try to be "generic". There is a discussion of endpoints at the top of the API doc to clarify.18:46
johnsomSome deploy with ports, some with just URL paths, etc.18:46
*** maciejjozefczyk has joined #openstack-lbaas19:02
*** maciejjozefczyk has quit IRC19:40
*** maciejjozefczyk has joined #openstack-lbaas19:41
*** Trevor_V has quit IRC19:41
*** gcheresh has joined #openstack-lbaas19:41
*** maciejjozefczyk has quit IRC20:06
*** TrevorV has joined #openstack-lbaas20:12
*** TrevorV has quit IRC20:18
*** gcheresh has quit IRC20:23
*** ivve has quit IRC21:08
*** zigo has quit IRC22:22
*** zigo has joined #openstack-lbaas22:31
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: Add skip_if_not_implemented to the service client  https://review.opendev.org/71400322:42
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: WIP: Adjust scenario tests for NotImplemented skip  https://review.opendev.org/71400423:00
openstackgerritMichael Johnson proposed openstack/octavia-tempest-plugin master: WIP: Adjust scenario tests for NotImplemented skip  https://review.opendev.org/71400423:02
*** spatel has joined #openstack-lbaas23:42
*** spatel has quit IRC23:46
« Wednesday, 2020-03-18 Index Friday, 2020-03-20 »

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