openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix syntax with gear unRegisterFunction() https://review.openstack.org/516403 | 03:03 |
---|---|---|
openstackgerrit | Rui Chen proposed openstack-infra/zuul feature/zuulv3: Use user home as work directory of executor https://review.openstack.org/516532 | 07:03 |
*** hashar has joined #zuul | 08:46 | |
*** electrofelix has joined #zuul | 09:20 | |
*** SpamapS has quit IRC | 09:36 | |
*** SpamapS has joined #zuul | 09:40 | |
*** sambetts|afk is now known as sambetts | 09:55 | |
*** xhku_ has joined #zuul | 10:01 | |
*** fbouliane has quit IRC | 10:01 | |
*** timrc has quit IRC | 10:01 | |
*** timrc has joined #zuul | 10:02 | |
*** tobiash has quit IRC | 10:12 | |
*** tobiash has joined #zuul | 10:15 | |
*** hashar is now known as hasharLunch | 11:12 | |
*** hasharLunch is now known as hashar | 13:09 | |
*** dkranz has joined #zuul | 13:21 | |
leifmadsen | jeblair: what time works for you tomorrow? | 14:13 |
jeblair | leifmadsen: how is 1500 utc (8am pacific, 11 eastern)? | 14:48 |
leifmadsen | jeblair: perfect | 15:15 |
leifmadsen | I'll schedule it | 15:15 |
jeblair | leifmadsen: thanks! | 15:15 |
leifmadsen | jeblair: I'll block it off for... 3 hours, but we aren't tied to using the whole thing | 15:15 |
*** bhavik1 has joined #zuul | 16:20 | |
*** bhavik1 has quit IRC | 16:25 | |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Increase github delay to 10 seconds https://review.openstack.org/515812 | 16:25 |
dmsimard | jeblair: not a fire but I've noticed disparity between ara versions on executors so it might not update itself automatically as we thought. I've seen at the very least 0.14.0, 0.14.2 and 0.14.4 ran today. I've added it to zuulv3-issues as "ARA (and other things?) are not kept up to date across the different executors" | 17:00 |
jeblair | here's an idea: have the merger perform an fsck on each git repo the first time it uses it after startup, and delete it if it fails | 17:00 |
jeblair | (this would include the executor's internal merger) | 17:00 |
jeblair | dmsimard: hrm | 17:00 |
jeblair | dmsimard: pip3 freeze agrees with you | 17:01 |
jeblair | dmsimard: but let's take this to -infra | 17:02 |
dmsimard | jeblair: it's likely to affect other modules we might want to keep up to date | 17:02 |
*** hashar is now known as hasharAway | 17:03 | |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Use user home as work directory of executor https://review.openstack.org/516532 | 17:21 |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Check start time for wait_time key https://review.openstack.org/516465 | 17:21 |
SpamapS | Ok, so I've now written the "extract all the projects' src dirs into their own variables" role 3 times. I think it may be time to add a dictionary by project to the zuul dict. | 18:07 |
clarkb | SpamapS: I think that is in progress but has been reverted a couple times? | 18:08 |
SpamapS | It can just be zuul.project_srcs or something. Doesn't have to change zuul.projects | 18:09 |
dmsimard | yeah we've been struggling for the projects to dict thing | 18:09 |
SpamapS | is there an unseen gotcha or just too busy? | 18:10 |
dmsimard | We've had to do two reverts so far, the first one was something obvious we missed during code review.. the second one is nasty, something around the ordering of the dict not being guaranteed | 18:10 |
dmsimard | let me pull up eavesdrop.. | 18:11 |
dmsimard | SpamapS: http://eavesdrop.openstack.org/irclogs/%23openstack-infra/%23openstack-infra.2017-10-29.log.html#t2017-10-29T17:17:10 | 18:12 |
dmsimard | it tried to clone oslo.serialization from branch feature/1.0 and feature/1.0 was an ara branch | 18:12 |
dmsimard | jeblair provided a summary in https://review.openstack.org/#/c/516117/ | 18:13 |
openstackgerrit | David Moreau Simard proposed openstack-infra/zuul-jobs master: Authorize the multi-node-bridge network in iptables if there's one https://review.openstack.org/516757 | 18:15 |
SpamapS | I think the list is important for some things | 18:15 |
dmsimard | SpamapS: not as far as we could tell | 18:15 |
SpamapS | I certainly like the fact that I can get the order in which the job user has requested projects, no? | 18:16 |
dmsimard | SpamapS: well, other than adjusting the different roles that expected a list | 18:16 |
SpamapS | oh so why break the interface? | 18:16 |
SpamapS | just create a new dict for things that need the dict? | 18:16 |
dmsimard | SpamapS: there's a short deprecation window, there is zuul._projects and zuul.projects | 18:16 |
SpamapS | -1 from me | 18:17 |
SpamapS | it's out there and in use.. and we're at least trying not to break people I thought. | 18:17 |
jeblair | SpamapS: that's after we relase v3 | 18:17 |
jeblair | SpamapS: we absolutely need to be able to change things while we're still in development | 18:17 |
SpamapS | yeah I'm down for changing things when it's necessary. I don't think this is, but I haven't seen the arguments, so I could be uninformed. | 18:18 |
jeblair | SpamapS: we only need one copy of this data structure. it can be very large. | 18:18 |
jeblair | it quite simply should have been a dict to start with. so we should fix it. | 18:19 |
SpamapS | Ok, I thought there was some important ordering being communicated. | 18:19 |
SpamapS | Sounds like it's just "we made everybody process it as a list" | 18:19 |
jeblair | SpamapS: no, as a list it's unordered | 18:19 |
jeblair | internally it's a set | 18:19 |
jeblair | so it goes set -> list currently | 18:19 |
SpamapS | doh | 18:20 |
SpamapS | glad it's already considered fully. | 18:20 |
SpamapS | I for one welcome our new dict overlords then. ;) | 18:20 |
jeblair | SpamapS: you convinced me :) | 18:20 |
SpamapS | Yeah I kind of remember asking for it. :) | 18:20 |
SpamapS | I just didn't know what would become of that request. | 18:20 |
jeblair | so yeah, i think the only gotcha remaining is our use of it in zuul-jobs where we need to learn how to get "with_dict" and "with_together" to play nicely together | 18:20 |
jeblair | SpamapS: so if you want to start using '_projects' now, you can start using it as a dict | 18:21 |
SpamapS | ooo a new toy | 18:21 |
* SpamapS has never used with_together | 18:21 | |
jeblair | SpamapS: then it'll be a one-char change once we switch .projects over to it | 18:21 |
SpamapS | jeblair: ack. Well I've consolidated all of my processing of zuul.projects into one role now, so it won't hit me too hard. | 18:22 |
jeblair | SpamapS: cool. i'd recommend going ahead and trying out zuul._projects when you have a sec; you'll either be ahead of the curve or find something else we need to fix before we make the switch :) | 18:23 |
SpamapS | jeblair: I have an interesting thing going on with zuul-jobs that is related to this btw. | 18:25 |
SpamapS | jeblair: since I've mirrored zuul-jobs in our GHE, I went ahead and added a few basic syntax check jobs to my local zuul-jobs repo... | 18:25 |
openstackgerrit | David Moreau Simard proposed openstack-infra/zuul-jobs master: Authorize the multi-node-bridge network in iptables if there's one https://review.openstack.org/516757 | 18:26 |
SpamapS | jeblair: but I _kind of_ want to run like, _every_ job that touches zuul-jobs. I wonder if there's a way to have like, a mode where I can say "walk through this pipeline in the layout and add any job that touches this project to this project." | 18:26 |
*** sambetts is now known as sambetts|afk | 18:34 | |
jeblair | SpamapS: hrm, interesting; i'll cipher on that | 18:49 |
SpamapS | jeblair: it may be too heavy.. but it would be a nice thing to make sure one doesn't break any jobs on changes to central repos. | 18:57 |
*** electrofelix has quit IRC | 19:25 | |
*** hasharAway is now known as hashar | 19:38 | |
*** dkranz has quit IRC | 20:43 | |
openstackgerrit | David Moreau Simard proposed openstack-infra/zuul-jobs master: Persist iptables rules https://review.openstack.org/513943 | 21:39 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: WIP: failing test for reconfiguration at failed head https://review.openstack.org/516799 | 23:41 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!