Tuesday, 2019-10-01

*** jamesmcarthur has quit IRC00:08
*** goldyfruit_ has joined #openstack-infra00:14
*** jamesmcarthur has joined #openstack-infra00:17
*** weshay has joined #openstack-infra00:24
*** yamamoto has quit IRC00:24
*** markvoelker has quit IRC00:26
*** gyee has quit IRC00:27
*** yamamoto has joined #openstack-infra00:28
*** smarcet has quit IRC00:28
*** jamesmcarthur has quit IRC00:31
*** jamesmcarthur has joined #openstack-infra00:32
*** smarcet has joined #openstack-infra00:33
*** smarcet has quit IRC00:37
*** yamamoto has quit IRC00:43
*** yamamoto has joined #openstack-infra00:44
*** yamamoto has quit IRC00:49
*** yamamoto has joined #openstack-infra00:55
*** jamesmcarthur has quit IRC01:14
*** jamesmcarthur has joined #openstack-infra01:17
*** ianychoi has quit IRC01:17
*** smarcet has joined #openstack-infra01:18
*** ianychoi has joined #openstack-infra01:18
*** dave-mccowan has joined #openstack-infra01:21
*** jamesmcarthur has quit IRC01:22
*** igordc has quit IRC01:26
*** jamesmcarthur has joined #openstack-infra01:40
*** auristor has quit IRC01:41
*** auristor has joined #openstack-infra01:53
*** hongbin has joined #openstack-infra01:55
*** jamesmcarthur has quit IRC02:06
*** jamesmcarthur has joined #openstack-infra02:07
*** ykarel|away has joined #openstack-infra02:12
*** yamamoto has quit IRC02:14
*** prometheanfire has quit IRC02:16
*** prometheanfire has joined #openstack-infra02:16
*** yamamoto has joined #openstack-infra02:17
*** hongbin has quit IRC02:33
*** roman_g has quit IRC02:33
*** hongbin has joined #openstack-infra02:34
*** hwoarang has quit IRC02:38
*** hwoarang has joined #openstack-infra02:48
*** serim has quit IRC02:50
*** yamamoto has quit IRC02:59
*** dave-mccowan has quit IRC03:04
*** psachin has joined #openstack-infra03:12
*** raukadah is now known as chandankumar03:23
*** spsurya has joined #openstack-infra03:29
openstackgerritIan Wienand proposed openstack/diskimage-builder master: yum-minimal: Don't install yum, install libcurl  https://review.opendev.org/68446103:30
openstackgerritIan Wienand proposed openstack/diskimage-builder master: pip-and-virtualenv : deprecate source for CentOS 8, new variables  https://review.opendev.org/68446203:30
openstackgerritIan Wienand proposed openstack/diskimage-builder master: simple-init: Use wrappers to call pip for glean install  https://review.opendev.org/68564303:30
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Fix networking for CentOS 8  https://review.opendev.org/68451303:30
openstackgerritIan Wienand proposed openstack/diskimage-builder master: CentOS 8 minimal testing and support  https://review.opendev.org/68486003:30
openstackgerritIan Wienand proposed openstack/diskimage-builder master: WIP: CentOS 8 support  https://review.opendev.org/68430803:30
*** gmann_afk is now known as gmann03:33
*** markvoelker has joined #openstack-infra03:41
*** markvoelker has quit IRC03:42
*** markvoelker has joined #openstack-infra03:43
*** yamamoto has joined #openstack-infra03:44
*** yamamoto has quit IRC03:49
*** hongbin has quit IRC03:56
*** udesale has joined #openstack-infra03:57
*** igordc has joined #openstack-infra03:58
*** ianychoi has quit IRC03:59
*** ianychoi has joined #openstack-infra04:00
*** whoami-rajat has joined #openstack-infra04:04
*** kjackal has joined #openstack-infra04:09
*** smarcet has quit IRC04:11
*** yamamoto has joined #openstack-infra04:15
*** ricolin has joined #openstack-infra04:16
*** rfolco has quit IRC04:17
*** yamamoto has quit IRC04:21
*** surpatil has joined #openstack-infra04:26
*** ykarel|away has quit IRC04:27
*** yamamoto has joined #openstack-infra04:28
*** odicha has joined #openstack-infra04:30
*** jamesmcarthur has quit IRC04:31
*** jamesmcarthur has joined #openstack-infra04:33
AJaegerconfig-core, please review https://review.opendev.org/679806 https://review.opendev.org/684374 and https://review.opendev.org/68127004:34
*** spsurya has quit IRC04:37
*** spsurya has joined #openstack-infra04:38
*** weshay has quit IRC04:40
*** ykarel|away has joined #openstack-infra04:48
*** yamamoto_ has joined #openstack-infra04:49
*** ykarel|away is now known as ykarel04:51
*** yamamoto has quit IRC04:51
*** pcaruana has joined #openstack-infra05:01
*** ianychoi has quit IRC05:03
*** ianychoi has joined #openstack-infra05:05
*** jaosorior has joined #openstack-infra05:05
*** dpawlik has joined #openstack-infra05:07
*** jbadiapa has joined #openstack-infra05:13
*** yamamoto_ has quit IRC05:16
*** yamamoto has joined #openstack-infra05:16
*** jamesmcarthur has quit IRC05:17
*** eernst has joined #openstack-infra05:19
*** xenos76 has joined #openstack-infra05:20
*** ianychoi has quit IRC05:27
*** ianychoi has joined #openstack-infra05:29
*** markvoelker has quit IRC05:47
*** jamesmcarthur has joined #openstack-infra05:47
*** igordc has quit IRC05:52
*** yamamoto has quit IRC05:57
*** ianychoi has quit IRC06:00
*** ianychoi has joined #openstack-infra06:01
openstackgerritIan Wienand proposed openstack/diskimage-builder master: yum-minimal : update mirrors for Centos 8  https://review.opendev.org/68443906:02
openstackgerritIan Wienand proposed openstack/diskimage-builder master: dib-python : handle centos 8  https://review.opendev.org/68444106:02
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Update locales for Centos 8  https://review.opendev.org/68444206:02
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Update redhat-common pkg-map for centos 8  https://review.opendev.org/68444906:02
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Add environment switch for centos8 to use dnf  https://review.opendev.org/68445906:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Use $YUM instead of direct calls in more places  https://review.opendev.org/68446006:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: yum-minimal: Don't install yum, install libcurl  https://review.opendev.org/68446106:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: pip-and-virtualenv : deprecate source for CentOS 8, new variables  https://review.opendev.org/68446206:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: simple-init: Use wrappers to call pip for glean install  https://review.opendev.org/68564306:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Fix networking for CentOS 8  https://review.opendev.org/68451306:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: CentOS 8 minimal testing and support  https://review.opendev.org/68486006:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: WIP: CentOS 8 support  https://review.opendev.org/68430806:03
openstackgerritIan Wienand proposed openstack/diskimage-builder master: Remove "failovermethod=priority" for Fedora (dnf)  https://review.opendev.org/68587606:03
openstackgerritMerged openstack/project-config master: Remove charm-neutron-api-genericswitch from infra  https://review.opendev.org/68127006:04
rm_workyeah I think Shade used to do some intelligent guessing about region name that the SDK connection doesn't do >_>06:07
rm_workwill be interesting to see if this is really the issue tho, because the connect URL looked correct and I'm not sure why lack of region would cause the certificate validation to fail :/06:09
*** surpatil has quit IRC06:12
*** surpatil has joined #openstack-infra06:13
yoctozeptofungi, clarkb: thx for handling :-)06:16
*** jamesmcarthur has quit IRC06:20
*** jbadiapa has quit IRC06:22
*** pgaxatte has joined #openstack-infra06:23
*** jamesmcarthur has joined #openstack-infra06:28
*** jtomasek has joined #openstack-infra06:29
*** lpetrut has joined #openstack-infra06:31
*** yamamoto has joined #openstack-infra06:35
*** ianychoi has quit IRC06:40
*** ianychoi has joined #openstack-infra06:42
*** defionscode has quit IRC06:43
openstackgerritIan Wienand proposed zuul/nodepool master: dib-functional : allow extra elements to be passed  https://review.opendev.org/68588406:44
*** iurygregory has joined #openstack-infra06:47
*** yamamoto has quit IRC06:47
openstackgerritMerged opendev/system-config master: logrotate: better rotation options  https://review.opendev.org/68278606:49
openstackgerritIan Wienand proposed openstack/diskimage-builder master: [wip] use mirrors for functional test builds  https://review.opendev.org/68588506:55
*** hongbin has joined #openstack-infra06:56
*** slaweq has joined #openstack-infra06:57
ianwUnknown configuration error  ... huh06:57
openstackgerritIan Wienand proposed openstack/diskimage-builder master: [wip] use mirrors for functional test builds  https://review.opendev.org/68588507:00
*** hongbin has quit IRC07:00
openstackgerritIan Wienand proposed openstack/diskimage-builder master: [wip] use mirrors for functional test builds  https://review.opendev.org/68588507:02
*** rpittau|afk is now known as rpittau07:04
*** jamesmcarthur has quit IRC07:04
*** zbr is now known as zbr|ruck07:08
*** xek has joined #openstack-infra07:12
*** gregoryo has joined #openstack-infra07:14
*** tesseract has joined #openstack-infra07:16
*** tosky has joined #openstack-infra07:16
*** apetrich has joined #openstack-infra07:17
*** dtantsur|afk is now known as dtantsur07:23
*** kopecmartin|off is now known as kopecmartin07:24
*** jbadiapa has joined #openstack-infra07:27
*** soniya29 has joined #openstack-infra07:30
*** admcleod has quit IRC07:34
*** jpena|off is now known as jpena07:43
*** ralonsoh has joined #openstack-infra07:45
*** pkopec has joined #openstack-infra07:46
*** admcleod has joined #openstack-infra07:47
*** markvoelker has joined #openstack-infra07:47
*** ykarel is now known as ykarel|lunch07:48
*** markvoelker has quit IRC07:51
*** denaitre has quit IRC07:58
*** Jeffrey4l has quit IRC07:59
*** Jeffrey4l has joined #openstack-infra08:00
*** ociuhandu has joined #openstack-infra08:02
*** apetrich has quit IRC08:02
*** rcernin has quit IRC08:02
*** rcernin has joined #openstack-infra08:03
*** jbadiapa has quit IRC08:03
*** tkajinam has quit IRC08:04
*** jbadiapa has joined #openstack-infra08:04
*** apetrich has joined #openstack-infra08:06
openstackgerritSlawek Kaplonski proposed opendev/irc-meetings master: Update chairs for various Neutron meetings  https://review.opendev.org/68589308:13
*** surpatil has quit IRC08:13
*** jaosorior has quit IRC08:16
*** dchen has quit IRC08:18
*** ociuhandu has quit IRC08:21
*** ociuhandu has joined #openstack-infra08:22
*** yolanda has joined #openstack-infra08:23
*** yamamoto has joined #openstack-infra08:26
*** e0ne has joined #openstack-infra08:29
mordredrm_work: it should be identical code08:31
mordredalso - region_name processing happens after certificate validation - it's about filtering things returned in teh catalog08:32
mordred(unless you have a region-name templated auth url)08:32
*** derekh has joined #openstack-infra08:32
*** xenos76 has quit IRC08:33
mordredcorvus: I think adding gcc and friends to python-builder seems like a fine idea. it *could* mask incomplete bindep files, but I don't think that's it's job08:36
*** surpatil has joined #openstack-infra08:37
*** ykarel|lunch is now known as ykarel08:54
AJaegerconfig-core, please review https://review.opendev.org/679806 and https://review.opendev.org/68437408:59
*** jamesmcarthur has joined #openstack-infra09:01
*** ociuhandu has quit IRC09:02
*** jamesmcarthur has quit IRC09:06
*** gregoryo has quit IRC09:14
*** gfidente has joined #openstack-infra09:16
*** ociuhandu has joined #openstack-infra09:23
openstackgerritMerged openstack/project-config master: Add airshipui project  https://review.opendev.org/67980609:26
*** ociuhandu has quit IRC09:28
*** rfolco has joined #openstack-infra09:29
*** jbadiapa has quit IRC09:29
*** jbadiapa has joined #openstack-infra09:30
*** ricolin has quit IRC09:44
*** markvoelker has joined #openstack-infra09:48
*** markvoelker has quit IRC09:53
*** xenos76 has joined #openstack-infra09:54
*** yamamoto has quit IRC10:02
*** ociuhandu has joined #openstack-infra10:20
*** jaosorior has joined #openstack-infra10:24
*** udesale has quit IRC10:27
*** udesale has joined #openstack-infra10:28
*** jovial[m] has joined #openstack-infra10:35
jovial[m]I'm not sure if this is the right channel, but I can't login to the wiki via openid, I get: ` OpenID error10:37
jovial[m]An error occurred: an invalid token was found. `, any ideas?10:37
*** soniya29 has quit IRC10:37
fricklerjovial[m]: what openid did you use? it seems to work fine for me with my ubuntu one id10:39
openstackgerritMerged zuul/zuul master: Replace tox-py36 by tox-py37  https://review.opendev.org/68215810:47
jovial[m]frickler: I was also using ubuntu one. I don't currently have a username. I got as far as the "choose a username" page and wwhatever I enter there, I always get that same error.10:49
jovial[m]this is the URL that I finish on: https://wiki.openstack.org/wiki/Special:OpenIDLogin/ChooseName10:49
*** jbadiapa has quit IRC10:53
*** yamamoto has joined #openstack-infra11:00
*** ykarel is now known as ykarel|meeting11:02
fricklerjovial[m]: hmm, I have no idea, then. maybe some other infra-root can take a look later, as that server also seems to be so old that I can't access it.11:04
*** udesale has quit IRC11:05
*** udesale has joined #openstack-infra11:06
jovial[m]frickler: OK, thanks for looking into it.11:07
*** soniya29 has joined #openstack-infra11:11
*** dave-mccowan has joined #openstack-infra11:11
*** roman_g has joined #openstack-infra11:14
*** udesale has quit IRC11:17
*** ykarel|meeting is now known as ykarel11:20
*** hrw has joined #openstack-infra11:26
hrwgood morning11:26
hrwafter some zuul change all links there lead to http://zuul.openstack.org/review.opendev.org/XYZ rather to http://review.opendev.org/XYZ11:29
hrw<a href="review.opendev.org/XYZ">XYZ,A</a> in page code11:29
fricklerhrw: that is a known issue, sorry for the trouble. hopefully someone can fix this soon11:31
hrwfrickler: I may look just do not know where D:11:31
openstackgerritMerged opendev/irc-meetings master: Update chairs for various Neutron meetings  https://review.opendev.org/68589311:31
*** jpena is now known as jpena|lunch11:38
*** jcoufal_ has quit IRC11:40
*** weshay has joined #openstack-infra11:41
AJaegerhrw: I think that was discussed in #zuul last Friday, you might want to read logs via eavesdrop.o.o11:41
hrwAJaeger: thanks.11:42
*** hrw has left #openstack-infra11:47
AJaegermnaser: could you check the comment in https://review.opendev.org/#/c/683963/1/roles/tox/library/test_tox_install_sibling_packages.py, please?11:48
*** markvoelker has joined #openstack-infra11:49
*** markvoelker has quit IRC11:54
*** yamamoto has quit IRC11:58
*** ianychoi has quit IRC12:10
*** ianychoi has joined #openstack-infra12:13
*** georgk has quit IRC12:21
*** fdegir has quit IRC12:21
*** fdegir has joined #openstack-infra12:22
*** georgk has joined #openstack-infra12:22
*** lpetrut has quit IRC12:24
*** yamamoto has joined #openstack-infra12:25
fungijovial[m]: i've seen that on a handful of accounts, something about specific ubuntu one sso openids doesn't seem to work with the version of the mediawiki openid extension we've got (not sure if it would work in other versions of that extension or not)12:31
fungijovial[m]: i was able to recreate the problem once myself by creating a new ubuntu one account and trying to authenticate with that, but i couldn't tell what the difference was between that account and two other accounts i already had which were working12:32
*** ricolin has joined #openstack-infra12:32
jovial[m]fungi: is there a workaround?12:32
fungijovial[m]: unless someone (likely someone with a much better grasp of php than me) finds time to try and dig into the mediawiki openid extension and add some additional debugging, possibly also in collaboration with one of the ubuntu one sso maintainers, the only workaround i know about right now is to ask someone with a working wiki account to make edits for you12:33
*** trident has quit IRC12:34
fungiwe're in the process of rebuilding and upgrading wiki.openstack.org, which will have a newer version of that extension, so that might also just make the problem go away, but there's no current estimate for when that will get done either12:34
*** trident has joined #openstack-infra12:35
*** yamamoto has quit IRC12:37
*** kjackal has quit IRC12:37
*** jpena|lunch is now known as jpena12:38
*** georgk has quit IRC12:45
*** fdegir has quit IRC12:45
*** georgk has joined #openstack-infra12:45
*** fdegir has joined #openstack-infra12:45
*** smarcet has joined #openstack-infra12:45
*** kjackal has joined #openstack-infra12:45
*** kjackal has quit IRC12:46
*** kjackal has joined #openstack-infra12:49
AJaegerinfra-root, FYI, we have zuul errors regarding the neutron feature/graphql branch that release automation accidently re-created. Change https://review.opendev.org/685955 will avoid recreation - and then I'll ask for removal of the branch again.12:50
fungithanks AJaeger!12:53
*** Goneri has joined #openstack-infra12:56
*** jaosorior has quit IRC12:58
*** rcernin has quit IRC12:58
*** jaosorior has joined #openstack-infra13:02
*** jbadiapa has joined #openstack-infra13:06
*** yamamoto has joined #openstack-infra13:10
*** dpawlik has quit IRC13:12
*** dave-mccowan has quit IRC13:13
*** dave-mccowan has joined #openstack-infra13:14
*** mriedem has joined #openstack-infra13:15
*** jbadiapa has quit IRC13:16
*** yamamoto has quit IRC13:17
*** dave-mccowan has quit IRC13:19
*** trident has quit IRC13:22
*** trident has joined #openstack-infra13:23
*** soniya29 has quit IRC13:23
*** aaronsheffield has joined #openstack-infra13:25
*** dpawlik has joined #openstack-infra13:29
*** jcoufal has joined #openstack-infra13:31
openstackgerritKostyantyn Kalynovskyi proposed openstack/project-config master: Add airship/apis project  https://review.opendev.org/68437413:32
AJaegerinfra-root, 685955  is merged, could you delete the neutron branch feature/graphql again, please?13:39
*** bdodd has joined #openstack-infra13:41
*** ociuhandu has quit IRC13:42
*** michael-beaver has joined #openstack-infra13:42
*** ociuhandu has joined #openstack-infra13:43
*** sthussey has joined #openstack-infra13:46
*** ociuhandu has quit IRC13:48
*** eernst has quit IRC13:49
*** markvoelker has joined #openstack-infra13:50
*** jaosorior has quit IRC13:51
*** markvoelker has quit IRC13:55
fungiCode Review - Error: branch openstack/neutron,refs/heads/feature/graphql has open changes13:57
*** markvoelker has joined #openstack-infra13:57
fungii expect the branch creation triggered a .gitreview patch proposal13:57
fungisearching now13:57
fungiabandoning https://review.opendev.org/685768 and https://review.opendev.org/68576713:58
*** odicha has quit IRC13:59
fungi#status log deleted feature/graphql branch of openstack/neutron previously at 9ee628dcc0eb61418e86fee355add38a6b09fab9 per https://review.opendev.org/68595513:59
*** SurajPatil has joined #openstack-infra13:59
openstackstatusfungi: finished logging13:59
fungiAJaeger: ^ thanks!14:00
openstackgerritMerged openstack/project-config master: Add airship/apis project  https://review.opendev.org/68437414:00
*** markvoelker has quit IRC14:01
AJaegerthanks, fungi . slaweq FYI ^14:01
*** jcoufal_ has joined #openstack-infra14:02
*** odicha has joined #openstack-infra14:02
*** surpatil has quit IRC14:02
*** jcoufal has quit IRC14:05
*** mtreinish has joined #openstack-infra14:06
slaweqAJaeger: fungi thx14:06
*** mattmceuen has left #openstack-infra14:06
*** surpatil has joined #openstack-infra14:08
*** ociuhandu has joined #openstack-infra14:10
*** SurajPatil has quit IRC14:11
*** SurajPatil has joined #openstack-infra14:20
*** markvoelker has joined #openstack-infra14:22
*** surpatil has quit IRC14:22
*** markvoelker has quit IRC14:26
*** surpatil has joined #openstack-infra14:26
*** markvoelker has joined #openstack-infra14:27
*** SurajPatil has quit IRC14:29
*** smarcet has quit IRC14:31
openstackgerritJames E. Blair proposed zuul/zuul-registry master: Initial implementation  https://review.opendev.org/68395814:34
openstackgerritJames E. Blair proposed zuul/zuul-registry master: Fix container image build  https://review.opendev.org/68580814:34
donnydfungi: Have you heard any complaints lately about FN? Just wanting to make sure the reduction in test nodes actually worked14:39
*** lpetrut has joined #openstack-infra14:40
clarkbI havent heard any14:41
donnydclarkb: I got around to setting up some irc notifications and I haven't either. I think 65 is what FN is reliably do without issues.14:42
donnydWhich bums me out a little, I really wanted to hit 100.. maybe more tuning can get me there14:42
*** dpawlik has quit IRC14:43
*** lpetrut has quit IRC14:47
mordredcorvus: re: adding gcc/g++ to the python-builder image ... I think the easiest thing would be to just switch python-builder to being based on python and not python:slim. python has gcc/make installed in it already14:51
openstackgerritMonty Taylor proposed opendev/system-config master: Use python base image for python-builder  https://review.opendev.org/68598914:52
mordredcorvus: like that ^^14:52
fungidonnyd: did you manage to figure out what resource contention you were hitting when pushing above 65?14:54
fungior is that still unknown?14:55
*** jamesmcarthur has joined #openstack-infra14:55
donnydfungi: I am thinking it was CPU. I am not oversubscribed on memory, and when I measured IO (with fio) at full load it appeared as though nothing was even running. I feel pretty safe in saying that I cannot oversubscribe CPU's on FN due to their age14:57
donnydAt 65 CPU subscription is around 105%14:58
fungimakes sense, it's hard to cleanly timeslice on x86 architecture too14:58
*** jbadiapa has joined #openstack-infra14:58
corvusmordred: oh neat that's easy14:59
donnydI am sure there is more tuning I could do, but I don't want to push much further past standard configs because I worry about the risk of jobs then passing on FN that don't pass other places from the huge bump in IO15:00
donnydI guess I can just be happy with what I can get out of it.15:01
*** xenos76 has quit IRC15:04
*** jbadiapa has quit IRC15:05
donnydfungi: It would be cool if I could get churn metrics on how many jobs FN does per hour or something like that15:05
donnydI am curious to know if the IO helps in getting those jobs done faster or if the old CPU's balance out the faster IO speeds15:06
fungimgagne: i don't know if it's important, but we've gotten a couple of notifications today that "our" credit card is expiring shortly, and to log into the internap portal and update our billing details to avoid any service interruption15:06
fungi(one for the contact address of each of our two tenants there)15:07
fungidonnyd: i'm trying to think if we have anything like that from statsd/graphite15:08
openstackgerritJames E. Blair proposed zuul/zuul-registry master: Initial implementation  https://review.opendev.org/68395815:10
openstackgerritJames E. Blair proposed zuul/zuul-registry master: Fix container image build  https://review.opendev.org/68580815:10
openstackgerritFabien Boucher proposed zuul/zuul master: WIP - Gitlab - Basic handling of merge_requests event  https://review.opendev.org/68599015:12
*** lpetrut has joined #openstack-infra15:13
*** ociuhandu has quit IRC15:17
*** ociuhandu has joined #openstack-infra15:18
*** smarcet has joined #openstack-infra15:19
*** ianychoi has quit IRC15:20
*** iurygregory has quit IRC15:21
*** ianychoi has joined #openstack-infra15:23
*** ociuhandu has quit IRC15:24
*** prometheanfire has quit IRC15:24
*** ociuhandu has joined #openstack-infra15:25
*** prometheanfire has joined #openstack-infra15:25
*** ykarel is now known as ykarel|afk15:26
*** ociuhandu has quit IRC15:30
*** ociuhandu has joined #openstack-infra15:31
*** mattw4 has joined #openstack-infra15:32
*** tesseract has quit IRC15:33
*** smarcet has quit IRC15:33
*** gyee has joined #openstack-infra15:34
*** lpetrut has quit IRC15:35
*** ociuhandu has quit IRC15:36
corvusinfra-root: i went to sign into gerrit's gerrit using infra-root's account, and google has decided that it wants to send me an sms code before it will let me log in.  i'm not willing to associate that account with my personal number.  does infra-root have a cell phone we can use for this purpose, or are we stuck?15:38
mordredwow15:38
mordredI feel like it would be nice to have an sms something that could be accessed somehow by infra-root people rather than being associated with someone's phone15:39
mordredare there reasonably free/open solutions for such a thing?15:39
clarkbcorvus: I think we can set up a u2f key (or pair of keys instead)15:39
clarkbor totp15:39
corvusclarkb: i'm not sure that's an option15:39
clarkboh hrm15:39
corvusthis isn't 2fa -- this is 'we think you are a bot'15:40
clarkbah15:40
fungiin fairness, the account is for a bot15:40
corvusindeed15:40
clarkbthat might be something luca knows how to handle?15:40
fungimight also be worth asking the gerrit community if this is something they've already worked out?15:40
clarkbsince they run a ci system separate of the google hivemind right?15:40
corvusi would bet a nickel someone just put their phone number in15:41
fungior they manually create ci accounts internal to gerrit rather than put them through google auth?15:41
fungi(like we used to do)15:41
mordredgoogling for "open source service for receiving automated SMS" gets lots of results for *sending* SMS15:41
*** jpena is now known as jpena|brb15:42
donnydWhy not use google voice and then tie phone numbers to that one google voice account15:42
clarkbdonnyd: doesn't that just nest the problem? still need sms for the voice account?15:42
*** ociuhandu has joined #openstack-infra15:42
*** rpittau is now known as rpittau|afk15:42
donnydvoice does sms15:42
mordredyeah - we'd need to be able to log in to a google account to add a voice account15:42
clarkbdonnyd: right but to verify the voice acount...15:43
donnydYou will need a real phone15:43
clarkbright15:43
clarkbwhich is the pickle we are in right now anyway15:43
donnydLOL15:43
*** xek has quit IRC15:44
donnydI see, so nobody wants to tie their number to it... Maybe a SIP account with sms support would work then15:44
*** rh-jelabarre has joined #openstack-infra15:44
clarkbcan our asterisk server do sms?15:44
*** xek has joined #openstack-infra15:44
fungihah, the trunk for the pbx? ;)15:44
fungiindeed15:45
corvushttps://wiki.voip.ms/article/SMS15:45
donnydcorvus: that is what a couple peoples on reddit said too15:45
corvusthat's the service we use for pbx.openstack.org so should be easy to try15:46
corvusthere's a good chance that google might know about that and choose not to accept that number15:46
corvusbut maybe it's worth a try :)15:46
*** ociuhandu has quit IRC15:46
donnydwell if you can just get direct texts, and given asterisk can support sms you should be able to forward from there no?15:47
donnydhttps://wiki.asterisk.org/wiki/display/AST/SMS15:47
corvusapparently voip.ms can forward the sms to email, so no ast setup required15:47
corvusi'll give this a shot :)15:48
* mordred expects great things15:48
*** ociuhandu has joined #openstack-infra15:49
donnydoh nice, well then that may be problem solved15:49
* donnyd gets coffee and waits for result15:50
corvusfungi, clarkb: the email address for the account is not one that i have access to; do you?15:50
donnydall of these anti-bot protections make it difficult to be a bot these days15:51
corvusfungi, clarkb: oh, it looks like you can specify an email address for sms, so that shouldn't be an issue15:51
corvusdonnyd: tell me about it15:51
corvus(ftr, i did ask the gerrit folks how they would like this account created, and i did that)15:51
clarkbcorvus: if its scotts address due to billing we can work something out if necessary I bet. But if you can set it to infra-root@o.o even better15:52
corvuswow you can have sms's trigger http url.  could totally add an sms trigger for zuul.15:53
mordredcorvus: and a zuul sms reporter too15:55
mordredfirst python contract I ever did was writing snpp support ... would be an amusing reporter - and very mean to connect someone's phone to15:55
mordredmaybe for april fool's we could have zuul sms jbryce every time a patch merges...15:56
corvuseven better if we can get one of those premium rate sms numbers and monetize it :)15:57
fungii remember at an early sysadmin job where we were issued sms-capable cell phones, i built an sms/e-mail interface to perform ping/traceroute and reply with the results15:57
mordredcorvus: "you can recheck a zuul job via sms for a small cost"15:57
corvusour "watson i want to see you" moment has arrived: i sent an sms message, "Sup?", to our number and have received an email15:59
clarkbmordred: tie it to promotion and charge even more :)15:59
fungiheh, charge a messaging fee for rechecks ;)15:59
corvusthat seemed to be the appropriate first sms to send15:59
fungithe telephony equivalent of "hello world"16:00
corvus"This phone number cannot be used for verification."16:01
clarkbwell then16:01
fungishoot16:01
fungiwith as many bots as google slams our webservers, it's amusing they're prejudiced against anyone else's bots16:03
corvuswell, i already got the api password, so strictly speaking, i don't think we're blocked now as long as we can use the rest api for everything we need16:03
*** surpatil has quit IRC16:04
*** pgaxatte has quit IRC16:09
*** surpatil has joined #openstack-infra16:10
openstackgerritClark Boylan proposed zuul/zuul-jobs master: Revert "Use import_role on persistent-firewall"  https://review.opendev.org/68599616:11
*** ociuhandu has quit IRC16:11
*** SurajPatil has joined #openstack-infra16:13
*** surpatil has quit IRC16:16
*** e0ne has quit IRC16:18
*** jpena|brb is now known as jpena16:19
*** lpetrut has joined #openstack-infra16:22
corvusi approved https://review.opendev.org/682749 which is the next step in talking to gerrit16:23
corvuswhen that's in and reconfigured, we recheck https://review.opendev.org/68275816:24
donnydcorvus: the only thing that really worries me is if gerrit will actually renew some car warranty16:24
*** ociuhandu has joined #openstack-infra16:25
*** weshay_ has joined #openstack-infra16:25
mordreddonnyd: Yeah, but I'm sayin', that TruCoat, you don't get it and you get oxidization problems.16:26
*** weshay has quit IRC16:26
*** SurajPatil has quit IRC16:27
* donnyd hopes gerrit doesn't also need a new medical plan16:28
*** ociuhandu has quit IRC16:29
openstackgerritJimmy McArthur proposed zuul/zuul-website master: CSS fix for ul/li in FAQ  https://review.opendev.org/68600316:31
*** ricolin has quit IRC16:33
*** smarcet has joined #openstack-infra16:36
*** ykarel|afk is now known as ykarel|away16:36
*** dtantsur is now known as dtantsur|afk16:40
openstackgerritMerged opendev/system-config master: Use the gerrit driver for the googlesource connection  https://review.opendev.org/68274916:40
*** ykarel|away has quit IRC16:43
*** weshay_ is now known as weshay16:46
*** imacdonn has quit IRC16:48
*** gfidente is now known as gfidente|afk16:55
*** smarcet has quit IRC16:58
*** derekh has quit IRC17:01
*** armax has quit IRC17:06
toskyuh, which component does create the zuul user and its directory (/home/zuul)? In the devstack/tempest jobs there is a difference in permissions between ubuntu images (0755) and centos 7 ones (0700),17:07
toskyand the latter breaks the jobs where the run-tempest role tries to access the tempest_blacklist file shipped by a tempest plugin (so somewhere under /home/zuul)17:08
toskyI'd say that 0711 would be a reasonable compromise, but I'm not sure where that value should be changed17:08
clarkbtosky: its part of our image build process with dib17:08
clarkbthere is a zuul element /me finds it17:08
clarkbtosky: that implies to me that centos and ubuntu have different modmasks17:09
clarkbtosky: https://opendev.org/openstack/project-config/src/branch/master/nodepool/elements/zuul-worker/install.d/60-zuul-worker17:10
toskyyep17:10
toskyI can see the different default permissions from a quick test on Fedora (which should have the same CentOS settings) and Debian (for Ubuntu, unless they changed)17:12
toskydo you think that setting the value to 0711 would be acceptable (so restricting it a bit on Ubuntu), or should I try to just set it for CentOS/RHEL/Fedora where permissions are more restricted?17:14
*** yamamoto has joined #openstack-infra17:15
*** jpena is now known as jpena|off17:16
*** xenos76 has joined #openstack-infra17:19
*** yamamoto has quit IRC17:19
AJaegerclarkb: ianw flagged https://review.opendev.org/683431 for PTL review if you want to tackle it at some point. it adds a special nested label for VMs.17:20
clarkbtosky: we probably need to set the modmask otherwise your subdirs with the git repos will be restricted too17:20
*** andreww has joined #openstack-infra17:21
clarkbAJaeger: thanks17:21
AJaegerconfig-core, any opinion on https://review.opendev.org/683028 (add a I'm not a robot voting option)? I'm torn, will +2 tomorrow probably...17:21
*** ykarel|away has joined #openstack-infra17:22
toskyclarkb: from what I've seen from some tests, that should be fine17:22
clarkbtosky: isn't the issue that the tempest user can't read from a git repo under the homedir? that will persist if you don't change the modmask17:22
clarkbtosky: you can open the top level dir but if the git repo isn't navigable it will still fail17:23
toskyclarkb: this is what I see right now: http://paste.openstack.org/show/780566/17:23
clarkbif we change the modmask then all new dirs will have the correct perms17:23
*** xarses_ has quit IRC17:23
clarkboh interesting so its the top level dir only that has a problem and modmask is fine otherwise17:24
toskyI understand the concerns, it's just that I'd prefer to touch as many directories as possible; I still remember how much time the patch which restricted the files touched by devstack took :)17:24
*** kjackal has quit IRC17:25
toskythe less likely is that someone starts shouting tomorrow and reverts the patch, the better I feel :)17:25
clarkbtosky: one option if you are wary about that is to have the job update the homedir perm early17:25
clarkbthen eg tripleo will be left untouched17:26
toskyactually tripleo jobs shouldn't suffer, as their permissions will increase17:26
toskyI'm more concerned on restricting them on ubuntu17:26
clarkbwell tripleo has explicitly complained about too permissive perms on dirs in jobs17:27
clarkbwhich is why I mentioned them17:27
toskyoh17:27
clarkband you don't necessarily have to remove perms on ubuntu17:27
clarkbyou can use relative perm adds like +x17:27
*** ralonsoh has quit IRC17:27
toskythat too17:27
toskyI just though that this more about a default zuul "thing", so fixing it in zuul would have been better17:28
clarkbya I'll have to think about it a bit more. My initial impression is that this would be a good default, but I've been wrong about that in the past :)17:29
toskyI vaguely rememeber some of the discussions about the permissions, maybe I can raise them with the people who complained17:30
toskydo you remember who was involved in the those discussions from the tripleo side?17:30
clarkbtosky: zbr|ruck17:31
zbr|ruckhi. i was about to leave.17:31
toskyzbr|ruck: oh, nothing urgent; when you can, please check the backlog17:31
*** lpetrut has quit IRC17:31
toskythe tl;dr question is: in the light of the previous discussions about permissions, would 0711 be acceptable for /home/zuul instead of the current centos default of 0700 ?17:32
zbr|ruckshortly is not a tripleo issue, is ansible default security model: not loading its config from folders which are world-writeable17:32
zbr|rucksomething that I find resonable17:32
pabelangerclarkb: tosky: I mean, the default under fedora is 0700 for homedirectory, so changing it across the board would be different in what users have. That said, I also think the idea of a pre-run jobs, is also good, for users that want to relax it17:33
toskypabelanger: in general, does it mean that any customizations applied to the images used by zuul should not diverge from the default content of the official cloud images?17:34
*** e0ne has joined #openstack-infra17:35
pabelangertosky: no, but as we diverge things, it makes them more different then what users get out of box17:35
pabelangerif job or app needs specific change, maybe better to also include that in the process of install or documentation over fixing in just CI images17:35
*** igordc has joined #openstack-infra17:37
toskyI guess I can add a task into the devstack.git, playbooks/pre.yaml, pre_tasks: section17:38
*** jamesmcarthur has quit IRC17:38
zbr|rucki kinda agreee with pabelanger that we should aim to keep systems as close to default/supported status whenever this is possible.17:41
toskythanks all17:42
*** armax has joined #openstack-infra17:47
*** jamesmcarthur has joined #openstack-infra17:48
*** kjackal has joined #openstack-infra17:51
toskyand I guess I can safely use ansible_user_dir instead of hardcoding /home/zuul17:53
corvusi'm running zuul-scheduler full-reconfigure to pick up the googlesource connection change17:58
*** jamesmcarthur has quit IRC17:59
*** jamesmcarthur has joined #openstack-infra17:59
*** Goneri has quit IRC17:59
*** xenos76 has quit IRC18:01
*** whoami-rajat has quit IRC18:02
corvusthat's done18:04
*** jamesmcarthur has quit IRC18:04
AJaegercorvus: is the Zuul Web UI link to changes issue also fixed?18:06
corvusoh, hrm, it looks like full-reconfigure doesn't restart the drivers, so this will need a full restart18:07
corvusAJaeger: i'm not aware of a change to fix that18:08
AJaeger;( ok18:08
corvusi'll look into doing a full restart after lunch+meeting18:09
*** psachin has quit IRC18:10
*** Goneri has joined #openstack-infra18:11
*** psachin has joined #openstack-infra18:12
*** smarcet has joined #openstack-infra18:15
*** xenos76 has joined #openstack-infra18:15
*** e0ne has quit IRC18:17
*** psachin has quit IRC18:17
*** jamesmcarthur has joined #openstack-infra18:19
*** jamesmcarthur has quit IRC18:23
*** harlowja has joined #openstack-infra18:32
*** Goneri has quit IRC18:34
*** ociuhandu has joined #openstack-infra18:34
*** ociuhandu has quit IRC18:37
*** ociuhandu has joined #openstack-infra18:39
openstackgerritMerged zuul/zuul-jobs master: Revert "Use import_role on persistent-firewall"  https://review.opendev.org/68599618:42
*** gfidente|afk has quit IRC18:45
*** ociuhandu has quit IRC18:46
*** ociuhandu has joined #openstack-infra18:46
*** jamesmcarthur has joined #openstack-infra18:47
*** Goneri has joined #openstack-infra18:48
*** jcoufal_ has quit IRC18:54
openstackgerritMerged zuul/zuul-jobs master: Allow upload-logs to toggle compression of console logs  https://review.opendev.org/56769618:56
fungicorvus: i think we may have collided on editing the secrets file. feel free to overwrite and i can readd mine when you're done. sorry!18:56
clarkbmeeting time in a couple minutes18:58
*** xenos76 has quit IRC19:00
corvusfungi: oh, i'm only reading it19:00
corvusfungi: i have exited19:00
corvusfungi: maybe double check that your update is there? (i expect it to be)19:00
fungicorvus: yep, it's in there. thanks!19:00
*** smarcet has quit IRC19:06
*** ociuhandu has quit IRC19:07
*** smarcet has joined #openstack-infra19:08
*** e0ne has joined #openstack-infra19:08
*** spsurya has quit IRC19:09
AJaegermnaser: could you check the comment in https://review.opendev.org/#/c/683963/1/roles/tox/library/test_tox_install_sibling_packages.py, please?19:13
*** pcrews has quit IRC19:15
openstackgerritMohammed Naser proposed zuul/zuul-jobs master: Add unit tests for tox_instalL_sibling_packages  https://review.opendev.org/68396319:18
mnaserAJaeger: done :>19:18
AJaegerthanks, mnaser19:19
AJaegerianw: your comment got addressed ^19:20
*** markvoelker has quit IRC19:21
*** smarcet has quit IRC19:24
*** xenos76 has joined #openstack-infra19:39
openstackgerritMerged zuul/zuul-jobs master: Add unit tests for tox_instalL_sibling_packages  https://review.opendev.org/68396319:42
*** eernst has joined #openstack-infra19:42
*** eernst has quit IRC19:46
*** smarcet has joined #openstack-infra19:46
rm_workeugh nm, i'm back to thinking i may need a hold19:48
rm_workcould someone put a hold for me on 683285?19:48
clarkbrm_work: for what job?19:48
rm_workor possibly tell me what they think could cause this?19:49
rm_workhttps://zuul.opendev.org/t/openstack/build/b164ed07603247579293699a1d6e007f19:49
rm_workSSLError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed (_ssl.c:852)')19:49
rm_work(this may belong in the SDK channel but ... this is related to the hold)19:50
rm_workit's a slightly different way of authing the SDK connection, and suddenly the certs don't verify19:50
rm_workwhich seems odd to me and i can't repro locally19:50
clarkbdon't use https against a mirror which only has http?19:50
rm_workso, it was JUST connecting19:50
rm_workif you look at the full log, this is the third run against the cloud19:50
rm_worki didn't change the URL19:51
rm_workat least I don't think I did?19:51
rm_workpreviously it was coming from the devstack supplied clouds.yaml so i guess possibly i'm using the wrong URL now? but this wasn't something I changed (the code has always been using that)19:51
rm_workand I would think the devstack would use HTTPS? does it not?19:52
rm_workI can try removing the S19:52
clarkboh what that isn't the mirror for some reason I read it as trying to fetch from the mirror19:52
clarkbits the localhost running openstack services19:53
rm_workyes19:53
clarkbit may not listen on the public address19:53
rm_workbut now that you say that, maybe it's possible that devstack didn't spin up for HTTPS?19:53
clarkbit may only listen on localhost or other private ip19:53
clarkbor ya no https19:53
rm_workhmm19:53
clarkbcheck the apache logs and config19:53
rm_workoh does it store the clouds.yaml19:54
*** fdegir has quit IRC19:54
rm_workthat'd be best to see19:54
rm_workyesss19:54
rm_workit does, sweet19:54
rm_workhmmmm19:54
rm_workAH19:54
rm_workthe clouds.yaml specifies a cacert19:54
*** fdegir has joined #openstack-infra19:54
rm_workwhich makes sense because it would be all snakeoil19:54
clarkbits not snakeoil, but it is a local CA19:55
rm_workkk this explains it19:55
rm_workso i need to give that cacert to the SDK19:55
rm_workthanks :D19:55
rm_work<319:55
rm_workdo you know how that clouds.yaml is built by devstack? is the path to that cacert in an ENV VAR somewhere/19:56
rm_workI could hardcode it but usually try to avoid that since i assume devstack puts EVERYTHING in an env var19:56
clarkbits in the devstack/tools/ dir as a script iirc19:56
rm_workor i could just do --insecure :D19:57
*** slaweq has quit IRC19:59
*** markvoelker has joined #openstack-infra20:00
rm_workAHA! it's $SSL_BUNDLE_FILE20:01
rm_work;)20:01
AJaegercorvus: is https://review.opendev.org/#/c/680696 ready now? "Switch to prepare-workspace-git role"20:01
rm_workthanks (again) for the pointer clarkb :)20:01
corvusAJaeger: i have not verified that we have new image builds since 68293420:03
corvusmaybe someone else has?20:03
*** eernst has joined #openstack-infra20:07
*** jamesmcarthur has quit IRC20:07
clarkbAJaeger: ianw left a commetn on https://review.opendev.org/#/c/683431/1 please let me know if that makes sense to you? Then maybe we can get donnyd logan- and mnaser to all +1 it20:08
donnydits already enabled and I think a couple of jobs are already using it, so happy FN can be of use20:10
*** kopecmartin is now known as kopecmartin|off20:11
*** eernst has quit IRC20:12
*** smarcet has quit IRC20:13
ianwclarkb: yeah, i didn't think we had a hard policy ... are you thinking more this is a "use this to get it stable and then jobs can dynamically switch" or we forever more keep the nodes types separate?20:15
donnydianw: I would think some of that is largely dependent on providers20:16
*** jamesmcarthur has joined #openstack-infra20:16
clarkbianw I hope that eventually most/all would end up enabling the feature if it is shown to be safe and reliable then we can remove the special subset20:16
*** ykarel|away has quit IRC20:17
clarkbas long as we have redundant clouds with the feature and everyone doesnt decide their jobs can only run there it should be fine to be separate for a while20:17
clarkbthen we can run targetted jobs on that label20:17
ianwyeah, that was my only concern, that jobs now evolve to only run there20:17
ianwrather than just happen to run faster there20:18
donnydafter its vetted, it would be really slick to have something that could parse if it was enabled or not, and then turn it on for jobs that run on providers that do have it20:19
corvusload is low and decreasing, so i'm going to do a full restart of zuul now20:19
*** ykarel|away has joined #openstack-infra20:20
clarkbianw ya I think that is the risk and its possible nested virt would stop working in anyor all cloudstoo20:20
ianwdonnyd: right, that's what i meant by "dynamically switch".  jobs should probe and decide20:20
clarkbso we should push people to not be dependent on it and instead start with targetted jobs20:20
ianwbut i understand that while you're trying to validate/test it all, it's a royal pain if you only get a nested capable host x% of the time (x < 100 :)20:20
donnydianw: well if you are sitting there waiting for the job to complete I would see how that could be frustrating.. but the jobs run now without it, so it shouldn't be much different. I would also think a random by product in FN's specific case is if the jobs that are already at the edge can move much faster.. I would be willing to bet I could over subscribe FN to do more jobs20:24
clarkbdonnyd: ya I think the point is to use a special label to prpve it works with targetted jobs that exercise the functionality20:25
clarkbthen ideally we can relax the rules more overall20:26
donnydyea that makes sense20:26
*** e0ne has quit IRC20:27
donnydShould ask sean-k-mooney if his NUMA jobs were using it20:27
clarkband if/when we have trouble with ut we know who to contact at these three clouds for cou details, kernel data, etc to file bugs with linux kernel20:29
*** kjackal has quit IRC20:29
fungiif some jobs run significantly faster with it, then i expect working out when to increase your job timeouts becomes a bit harder20:31
corvusi think there may be a problem with zuul in this configuration20:32
*** ykarel|away has quit IRC20:33
*** pcaruana has quit IRC20:33
corvusi'm going to manually remove the googlesource entry and restart again20:33
corvusor reconstrue it as a git driver20:34
donnydclarkb: I am still working on a way that doesn't significantly compromise my infra to push all log data public so you can see what is happening under the hood20:36
*** yamamoto has joined #openstack-infra20:37
corvusre-enqueuing now20:40
*** xek has quit IRC20:43
corvusoh i think i see what the problem was and it was me; i think i misnamed the hiera value for the password20:45
corvusi'll try to verify that was the only problem before i proceed20:46
corvusand maybe improve the error handling20:46
toskyuhm, have you noticed that the link to reviews from the zuul dashboard does not bring to the review (review.opendev.org/<nnnn>) but to zuul.openstack.org/review.opendev.org/<nnnn>?20:48
toskymaybe a relative link instead of an absolute one somewhere?20:48
*** pkopec has quit IRC20:49
fungiyes, it's known, though i don't know if anyone has gotten to the bottom of what changed to cause it (could be an update of a dependent lib), nor whether anyone has started work on fixing it20:49
toskyis it on a graphical-full-of-javascript-part-of-code?20:50
corvustosky: i suspect it's python20:51
clarkbshrews was looking at it when we were at ansiblefest iirc20:51
clarkbbut don't think it got tracked down20:51
corvusi also suspect i broke it20:51
corvushowever, it's like #4 on the list of things i may have broken and need to fix20:52
toskysure, I understand; it's just a minor annoyance20:53
Shrewsclarkb: corvus: fungi: on my last attempt to hunt it down, i tried to reproduce it locally. but it was not reproducible for me using HEAD of master branch.20:57
*** Goneri has quit IRC20:57
*** whoami-rajat has joined #openstack-infra20:58
fungiwell, the scheduler and web service were both restarted a few minutes ago, so presumably using ~head20:58
*** smarcet has joined #openstack-infra21:00
clarkbok the issue is in the status json the url change.url value doesn't have a https:// prefix. Its just review.opendev.org/$changenumber21:00
clarkbso the browser treats that as a relative path21:00
*** ociuhandu has joined #openstack-infra21:08
*** yamamoto has quit IRC21:08
corvusall right, the good news is that identified what i missed adding to the connection config.  the bad news is that in my local testing, when i configure it with our credentials, i get 403 unauthorized, which means that we *are* stuck on providing a phone number to google.21:11
*** yamamoto has joined #openstack-infra21:11
corvusour google friends are in germany, so we may need to wait until tomorrow to see what they say21:11
*** ociuhandu has quit IRC21:12
clarkbI htink it is possible that the url problem originates in https://review.opendev.org/#/c/681936/6/zuul/driver/gerrit/gerritmodel.py21:14
clarkbalmost like we parse ssh events a http21:15
corvusprivate "hiera" is corrected21:16
clarkbsince the http version drops the scheme in self.url there after parsing the url provided in data21:16
clarkbcorvus: ^ do you know why we don't pass through self.url as whatever data['url'] is in both cases?21:16
corvusclarkb: not offhand but i can look in a few minutes21:17
openstackgerritJames E. Blair proposed opendev/system-config master: Revert "Use the gerrit driver for the googlesource connection"  https://review.opendev.org/68604821:19
openstackgerritJames E. Blair proposed opendev/system-config master: Use the gerrit driver for the googlesource connection  https://review.opendev.org/68604921:19
corvusinfra-root: we should merge 686048 soon21:19
clarkbcorvus: ya that is what is happening. We create a session object for the review.opendev.org connection because we have a password for it (I'm guessing to report via http)21:20
clarkbbecause self.session is not none we create http gerrit change objects not ssh gerrit change objects and the http version of the objects remove the scheme from change.url21:20
clarkbI think we may be able to simply set self.url = data['url'] in the change object regardless of using http or ssh21:21
paladoxcorvus edwin is online21:25
corvuspaladox: yeah, i sent him a message, but it must be late so i won't be surprised if he politely ignores it until tomorrow :)21:26
paladoxoh21:26
*** openstackstatus has quit IRC21:32
*** openstackstatus has joined #openstack-infra21:34
*** ChanServ sets mode: +v openstackstatus21:34
*** odicha has quit IRC21:34
*** odicha has joined #openstack-infra21:35
openstackgerritClark Boylan proposed zuul/zuul master: Always use change url from gerrit api data  https://review.opendev.org/68605421:47
clarkbcorvus: ^ I went ahead and pushed a change to have a breadcrumb to follow21:48
corvusclarkb: thanks, i'll take a look at that now21:51
corvusclarkb: commented21:56
*** mriedem has quit IRC21:56
clarkbaha21:57
fungiso was this an outgrowth of the switch to use the rest api instead of ssh api?21:57
clarkbfungi: yes, which actually points to an oddity. I think the http parser is ebing used to parse ssh events21:57
fungineat!21:57
jrossersomething seems a bit brok at http://zuul.openstack.org/status for me21:58
clarkbfixing that likely requires more thought21:58
fungijrosser: what exactly?21:58
jrosserif i click one of the links in a job which should take me back to the gerrit page it tries to go to somewhere like this http://zuul.openstack.org/review.opendev.org/68100221:58
fungijrosser: yeah, that's the fix we were just talking about, though it's been this way for a week or two21:59
jrosseroh, no worries :)21:59
fungijrosser: clarkb started on a fix with 686054 just a few minutes ago21:59
corvusclarkb: aroo?  http parser ssh what?21:59
clarkbcorvus: the switch for deciding if the gerrit event should be handed as an ssh event or http event is whether or not we have an http session. Because our review.opendev.org connection config has a password we create a session for it, but only use that to report not get events. But we parse all the ssh events coming in as http events i think22:01
clarkbcorvus: the ssh parser shouldn't have this bug with url generation22:01
clarkbonly the http parser does and I think that is why we use it (the presence of a session for reporting)22:01
clarkbI also don't know if connection.baseurl will be an ssh:// or https:// url in this case22:02
corvusclarkb: nah -- think of this in terms of triggers, reporters, sources.  triggering actions for us use ssh.  but reporters and (with the recent change to support http more fully) sources use http when available.  when our event arrives over ssh, the source uses http to get information about the change.  that's what's happening here.22:02
clarkblooks like its always https based on the connection code22:02
corvusclarkb: baseurl is not set in our config file so it will be https://review.opendev.org22:03
clarkbcorvus: doesn't the ssh event have all of this info in it though?22:03
clarkbseems like we are doing an unnecassary round trip22:03
corvusclarkb: nope, it's missing a lot22:03
openstackgerritMerged opendev/system-config master: Revert "Use the gerrit driver for the googlesource connection"  https://review.opendev.org/68604822:03
corvusclarkb: it has info about the event, including some rudimentory information about the change, but for the full info we've always gone back to gerrit.  we used to do that over ssh, but now use http.22:04
clarkbI see22:05
corvusi expect we'll drop ssh source and reporter support in the not-too-distant-future for ease of maintenance.22:05
corvustrigger, of course, we'll keep :)22:05
openstackgerritClark Boylan proposed zuul/zuul master: Set url scheme on HTTP Gerrit events  https://review.opendev.org/68605422:13
clarkbthat better?22:13
*** aaronsheffield has quit IRC22:15
corvusclarkb: yep, though 'connection.baseurl' would have worked too22:15
corvusclarkb: thanks!22:16
*** dchen has joined #openstack-infra22:24
*** Goneri has joined #openstack-infra22:28
*** weshay has quit IRC22:29
*** xenos76 has quit IRC22:30
toskyoh, thanks!22:45
*** dciabrin_ has quit IRC22:45
*** jamesmcarthur has quit IRC22:48
*** Goneri has quit IRC22:48
*** smarcet has quit IRC22:49
*** tkajinam has joined #openstack-infra22:50
*** bdodd has quit IRC22:54
fungiclarkb: looks like the quickstart job for it failed out on a 404 during "TASK [Fetch log URL]"22:54
fungihttps://e1b7c5837d1435fd98ed-957e0323e7657520e3a7352c7d8ac713.ssl.cf5.rackcdn.com/686054/2/check/zuul-quick-start/ed703d6/job-output.txt22:54
fungihttps://e1b7c5837d1435fd98ed-957e0323e7657520e3a7352c7d8ac713.ssl.cf5.rackcdn.com/686054/2/check/zuul-quick-start/ed703d6/container_logs/logs.log22:55
*** bdodd has joined #openstack-infra22:55
clarkbhrm does that mean the test job failed?22:55
clarkbno executor log seems to show it succeeding22:56
*** yamamoto has quit IRC22:57
fungiyeah, executor log says the file in question was uploaded 3 seconds earlier if i'm reading it right22:57
fungithis is... weird22:57
clarkbat 22:48:34,57922:57
clarkb22:48:37 is when the apache 404 is logged. I agree seems like the file should be there22:58
clarkbmaybe an fs syncing issue22:58
*** weshay has joined #openstack-infra22:58
clarkbif these things are bindmounted across different containers?22:58
*** smarcet has joined #openstack-infra22:59
fungiuploaded as /srv/static/logs/2/2/1/check/testjob/062504d/job-output.txt and the request was for .../2/2/1/check/testjob/062504d/job-output.txt.gz22:59
clarkboh22:59
clarkbthis is that change that AJaeger was tlaking about iirc22:59
fungiwe merged a change recently to compress those unconditionally, right?22:59
fungior maybe other way around23:00
clarkbhttps://review.opendev.org/56769623:00
clarkbok I see what the fix is23:00
fungiso the yeah, i guess the quickstart exercise needs patching23:00
fungii suppose we don't run the quickstart job against zuul-jobs changes23:01
*** yamamoto has joined #openstack-infra23:01
fungiso didn't catch it before merging23:02
*** tosky has quit IRC23:02
openstackgerritClark Boylan proposed zuul/zuul master: Set url scheme on HTTP Gerrit events  https://review.opendev.org/68605423:02
openstackgerritClark Boylan proposed zuul/zuul master: Handle new default of uncompressed job-output  https://review.opendev.org/68607223:02
clarkbfungi: ^ I think that should do it23:03
*** weshay has quit IRC23:12
*** weshay has joined #openstack-infra23:15
*** jamesmcarthur has joined #openstack-infra23:32
*** smarcet has quit IRC23:33
*** rcernin has joined #openstack-infra23:33
*** bdodd has quit IRC23:37
*** markvoelker has quit IRC23:38
*** yamamoto has quit IRC23:39
*** yamamoto has joined #openstack-infra23:39
*** yamamoto has quit IRC23:43
*** markvoelker has joined #openstack-infra23:44
*** sthussey has quit IRC23:44
*** exsdev has quit IRC23:48
*** yamamoto has joined #openstack-infra23:48
*** exsdev has joined #openstack-infra23:48
*** smarcet has joined #openstack-infra23:49
openstackgerritIan Wienand proposed openstack/diskimage-builder master: [wip] use mirrors for functional test builds  https://review.opendev.org/68588523:51
*** yamamoto has quit IRC23:52
openstackgerritMerged zuul/zuul master: Handle new default of uncompressed job-output  https://review.opendev.org/68607223:54
*** yamamoto has joined #openstack-infra23:55

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