Wednesday, 2021-01-06

*** cloudnull has joined #opendev00:03
clarkbhas anyone had a chance to read the opendev project update for the annual report yet?00:06
clarkbI'm about to pass the zuul one over to the foundation to get feedback before things are due next week and would be great to get the opendev side in a similar spot00:06
corvusclarkb: looking now; btw, did i mention that my meetpad change was merged upstroam in jitsi?  (but last i looked they hadn't updated docker images)00:07
clarkbno I hadn't heard that00:07
corvushopefully soon we can drop our tiny fork (which is old)00:08
clarkbwe should be able to pull upstream for that component and still build our own image though? maybe its easiest to just switch to their image00:08
corvusyeah, we should switch to theirs; ours is only for that patch00:08
*** cloudnull has quit IRC00:10
corvusclarkb: opendev lgtm; though we might be able to call out the zuul image build pipeline stuff a bit more; maybe something about how we made it available to all projects and they can use depends-on and whatever00:12
clarkbthat is a good idea00:12
clarkbI'll try to work that in00:12
fungii briefly glimpsed at it but was hoping to take a closer look tomorrow00:13
corvusi'm not sure if we need to say the words "intermediate registry" but we could if that helps00:13
clarkbfungi: cool I can hold off on foundation feedback until you've had a chance to read it00:13
corvusfungi is foundation feedback :)00:13
fungiwell, i can also just count as "foundation feedback"00:13
fungiheh, indeed00:13
clarkb++00:14
clarkbcorvus: paragraph starting at line 67 tries to capture the image build stuff with a bit more detail if you can take another look00:22
*** tosky has quit IRC00:23
corvusclarkb: minor suggestion on line 69, take it or leave it.  otherwise, lgtm!00:35
clarkbdone and thanks00:36
*** DSpider has quit IRC01:13
*** cloudnull has joined #opendev01:42
*** diablo_rojo has quit IRC03:37
*** ysandeep|away is now known as ysandeep03:47
*** ykarel has joined #opendev04:53
*** whoami-rajat__ has joined #opendev05:41
*** slittle1 has quit IRC06:48
*** slittle1 has joined #opendev06:49
*** marios has joined #opendev06:57
*** prometheanfire has quit IRC07:14
*** sboyron has joined #opendev07:25
*** prometheanfire has joined #opendev07:29
*** rpittau|afk is now known as rpittau08:01
*** andrewbonney has joined #opendev08:23
*** hashar has joined #opendev08:42
*** ysandeep is now known as ysandeep|lunch08:46
*** elod_pto is now known as elod08:48
*** DSpider has joined #opendev09:33
*** ysandeep|lunch is now known as ysandeep09:40
*** slaweq has joined #opendev09:55
*** dtantsur|afk is now known as dtantsur10:35
*** slaweq has quit IRC10:38
*** hashar is now known as hasharLunch10:49
*** Eighth_Doctor has quit IRC11:20
*** mordred has quit IRC11:21
*** mordred has joined #opendev11:29
*** Eighth_Doctor has joined #opendev11:47
*** tosky has joined #opendev12:15
*** lpetrut has joined #opendev12:17
*** slaweq has joined #opendev12:17
*** ysandeep is now known as ysandeep|afk12:57
*** brinzhang0 has quit IRC13:12
*** slaweq has quit IRC13:15
openstackgerritSorin Sb├órnea proposed opendev/system-config master: Recognize LP urls for footer bugs  https://review.opendev.org/c/opendev/system-config/+/73796013:21
*** ysandeep|afk is now known as ysandeep13:33
dtantsurhi folks! are there any docs on how to use search in storyboard?13:58
fungidtantsur: the api or the webui?14:42
fungithe webui search widget is basically auto-completing what you type against a variety of different object types14:43
fungii don't know that it's particularly documented. the api has good docs though14:43
fungithe search options in the api are object-type-scoped, for example https://docs.openstack.org/infra/storyboard/webapi/v1.html#get--v1-stories-search14:45
dtantsurmm, okay, maybe I should use the API. the web UI doesn't seem usable (or I haven't figured it out)14:46
fungidtantsur: i click in the search field, my browser outlines it with a blue glow if it's active (i'm not sure what the significance of this is but i've noticed if i try to click in it while things are still loading i don't get that "blue glow" and then the typeahead searching doesn't seem to happen) and start typing some text. matches should (after a bit of a delay, could be a few seconds because the api can14:54
fungibe slow/overloaded at times) start to appear and then i can click on which one i want14:54
fungithere's a fix merged for a major api speedup but we haven't deployed it because we need to switch up how we're deploying the software and move to a newer distro with a version of python storyboard supports14:55
fungiif i click in the search field and don't get a blue border or it appears and then immediately disappears again, clicking some dead space on the page and then clicking back into the search field seems to solve it. not sure what's going on there14:57
fungisomething to do with javascript focus detection i guess14:57
dtantsurso, I tried inputting "ironic cleaning ramdisk". I got a "helpful" suggestion of "ironic cleaning ramdisk" :) I clicked on it and got everything in the world that has one of these words somewhere.14:57
dtantsurtry doing the same with google and site:storyboard.openstack.org and you'll get much more relevant results14:58
dtantsurI also could not figure out how to limit the search to only stories with a task on openstack/ironic14:58
fungiahh, yeah, right now it's just keyword matching, there is no query language. there's been discussion of redoing all the query interfaces and implementing something like a lucene parser15:00
dtantsurI see :(15:13
*** lpetrut has quit IRC15:27
*** slittle1 has quit IRC15:40
*** slittle1 has joined #opendev15:40
*** spotz has quit IRC15:44
*** mlavalle has joined #opendev16:01
*** ykarel has quit IRC16:30
openstackgerritMerged openstack/project-config master: Enable tripleo core members to change WIP flag  https://review.opendev.org/c/openstack/project-config/+/76582116:55
*** ysandeep is now known as ysandeep|away16:56
*** marios is now known as marios|out16:57
*** marios|out has quit IRC17:00
*** hasharLunch is now known as hashar17:02
clarkbI don't see any new leaked ansible processes on bridge which implies we've cleaned up the holiday migration set17:05
clarkbthere are 4 processes that I must've missed yesterday though which I'll clean up momentarily17:05
*** stzsch has quit IRC17:13
*** stzsch has joined #opendev17:18
*** rpittau is now known as rpittau|afk17:43
*** ttakenak has joined #opendev18:02
*** ttakenak has quit IRC18:03
openstackgerritClark Boylan proposed opendev/system-config master: Force run-gitea job to fail for human verification  https://review.opendev.org/c/opendev/system-config/+/76959818:14
clarkbinfra-root I have set an autohold on ^18:15
clarkbthat change is based on the gitea 1.13.1 change and it pushes in the local system-config into localhost then fails. This way we can hold the job then verify at least one repo has content and looks as expected18:16
clarkbI'm not entirely sure the https push will work eitehr and if that fails thats fine too we'll just have to manually update a repo once we have the held node18:19
*** dtantsur is now known as dtantsur|afk18:27
fungiweird... https://grafana.opendev.org/d/9XCNuphGk/zuul-status?orgId=1&from=now-6h&to=now shows that we briefly had a 13th zuul executor between 17:18 and 17:3218:40
fungistatsd/graphite error i guess?18:40
fungiyou can also see the merger count jumped from 20 to 21 for the same timeframe18:41
clarkbya I'm gonna guess that is an averaging thing18:41
fungii don't see any corresponding dips, so it's not like increment/decrement arrived out of order or anything18:42
*** andrewbonney has quit IRC18:56
clarkbok 769598 failed to do the push since ssl wasn't verified (we use a self signed cert)19:18
clarkbShould be able to get on that host then do the push against with GIT_SSL_NO_VERIFY set. I'll look into that a bit later today after lunch19:19
*** DSpider has quit IRC20:13
*** whoami-rajat__ has quit IRC20:21
prometheanfirecan I get a second view on https://review.opendev.org/768785 (glean) and possibly a relase once merged, gentoo glean is broken and it'd be nice to get the linked review in as well20:42
* fungi checks20:44
prometheanfirelol20:45
fungiahh, yeah, i already reviewed that so not helping20:45
prometheanfirewas going to mention you already looked, but didn't want to ping you about something you can't help with :P20:45
* fungi apologizes for being helpfully unhelpful20:46
prometheanfirefungi: you haven't looked at https://review.opendev.org/762339 :P20:50
clarkbprometheanfire: fungi I'm not sure the new code is correct either? /etc/hosts allows any whitespace delineation20:54
clarkbthis will only work if a single space is used20:54
clarkbI'm not sure how big of aconcern that is but that may be a reason to use a regex?20:55
prometheanfire /etc/hosts will be initialized via glean, so it'll be written out with a space, if modified after that it could be more of a concern, but not sure if that's in scope20:56
clarkbprometheanfire: for the magic mock does it just return '' for every read while iterating?20:58
clarkbor maybe returns one '' then eofs?20:58
clarkb(want to make sure I understand the test change there)20:58
prometheanfirethat's the way I understand it (1 '' then eofs)20:58
fungilike magic20:59
prometheanfireI am not the expert on mock, or really python in general20:59
clarkbya that makes more sense otherwise it would never end the iteration20:59
fungiyeah, honestly test mocks are black magic to me as well20:59
fungiwhich is part of why i didn't just single-core approve that change20:59
clarkbfungi: do you think the whitespace handling in the /etc/hosts line check is a problem or are we ok with matching what we write and if duplicates end up there due to user intervetion that is their problem?20:59
fungiclarkb: i think it's fine until we find out it isn't, and then we can fix it21:00
clarkbok21:00
fungii honestly didn't think about it21:00
prometheanfireright now it's writing a new set of host entries every reboot for me21:01
fungiit's not hard to switch to a re.match() or something though if needed21:01
clarkbI'm going to reboot to apply updates then dig up credentials and can approve it21:01
prometheanfirethanks21:01
*** sboyron has quit IRC21:02
clarkbone last question (just thinking out loud) r+ is like r + a modes right?21:15
clarkbit doesn't truncate and the write pointer is at eof?21:15
clarkbthe python docs don't really clarify that second bit, it does say it doesn't truncate which is probably good enough as order doesn't matter here I don't think21:15
clarkbbut I can test this really quickly21:15
clarkbya it does not append it prefixes21:16
clarkbfungi: prometheanfire does that matter?21:16
prometheanfireclarkb: I think r+ is more rw21:17
clarkbya I think you may want a+ ?21:18
clarkbto append rather than prefix21:18
clarkbtesting a+ seems to do what I would expect21:18
prometheanfireok, I'll push that up21:19
clarkbI don't think it truly matters but will help keep the glean stuff later in the file after any preexisting system stuff21:19
clarkbwhich will be nice for users21:19
prometheanfireyep21:20
openstackgerritMatthew Thode proposed opendev/glean master: add hosts entries and ssh keys only once  https://review.opendev.org/c/opendev/glean/+/76878521:20
clarkbprometheanfire: hrm actually a+ reads don't work21:24
clarkbugh21:24
clarkbthey read at eof too I think21:24
prometheanfirelol21:24
prometheanfireso, r+ then?21:24
clarkbno ithink you want a read handle and a write handle21:24
clarkbread from from beginning to end of file and if not present then append21:24
clarkber then append using another file object opened with a21:25
prometheanfirethat's what it's currently doing, I was hoping to use just one file object21:25
clarkbI don't think python's file objects do read and write pointers at different places? at least that seems to be the issue with a+21:25
prometheanfireright21:26
prometheanfirehow it's written now (with r+) is to do the writes at the end of file, I think21:27
fungiclarkb: worth noting, the code there iterates over the entire file and then writes, so it's in essence a+21:27
fungii did at least double-check that by hand21:27
clarkbfungi: oh ok ya I was missing that21:27
fungiis works as expected21:27
prometheanfireso... r+ then?21:27
fungithe file cursor is already at the eof when the write() call happens21:27
clarkbprometheanfire: ya the iteration to the end makes it an apend and r+ doesn't truncate so ya sorry tht was correct21:27
prometheanfirecool21:28
openstackgerritMatthew Thode proposed opendev/glean master: add hosts entries and ssh keys only once  https://review.opendev.org/c/opendev/glean/+/76878521:28
clarkbok +2'd sorry for that noise. fungi if you want to reapply your +2 then approve I'm happy with that21:30
fungisure21:34
*** hashar has quit IRC22:23
*** tosky has quit IRC23:51
openstackgerritMerged opendev/glean master: add hosts entries and ssh keys only once  https://review.opendev.org/c/opendev/glean/+/76878523:52

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!