| mordred | corvus: heh: Sending result: {"result": "ERROR", "error_detail": "Project ansible/ansible does not have the default branch master"} | 00:01 |
|---|---|---|
| clarkb | it doesnt? | 00:02 |
| mordred | no- they have no master branch and instead use devel | 00:02 |
| mordred | I've added an override-branch setting to the required-projects | 00:04 |
| clarkb | why do projects do that? | 00:04 |
| clarkb | that value isnt replicated on pushes its a git failing but means you should just use master | 00:04 |
| mordred | yah. I don't grok it myself | 00:05 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add job for testing against devel of ansible https://review.openstack.org/531934 | 00:06 |
| clarkb | it is an unfortunate git behavor but it is what it is | 00:08 |
| mordred | clarkb, corvus: http://zuulv3.openstack.org/stream.html?uuid=6e346ac9d4054654b7db8edabf45aa58&logfile=console.log <-- it's at least past the zuul config error - and I see github.com/ansible/ansible in the projects list | 00:09 |
| mordred | 2018-01-09 00:09:21.123535 | controller | Cloning into '/home/zuul/src/github.com/ansible/ansible'... | 00:09 |
| mordred | 2018-01-09 00:09:33.674617 | controller | ok: Item: Runtime: 0:00:12.462275 | 00:09 |
| mordred | PIDDLE | 00:14 |
| mordred | well - tox-install-siblings doesn't work with ansible ... patch coming | 00:15 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Use setup.py if we can't get the name from setup.cfg https://review.openstack.org/531936 | 00:24 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Add job for testing against devel of ansible https://review.openstack.org/531934 | 00:25 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Add /node-list to the webapp https://review.openstack.org/499969 | 01:54 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/nodepool feature/zuulv3: Add /label-list to the webapp https://review.openstack.org/500254 | 01:54 |
| *** haint has joined #zuul | 02:20 | |
| *** nguyentrihai has quit IRC | 02:24 | |
| *** harlowja has quit IRC | 02:32 | |
| *** rlandy_ has quit IRC | 02:37 | |
| *** rlandy|bbl is now known as rlandy | 02:38 | |
| *** rlandy has quit IRC | 02:38 | |
| *** xinliang has quit IRC | 02:43 | |
| *** xinliang has joined #zuul | 02:54 | |
| *** JasonCL has quit IRC | 03:03 | |
| *** JasonCL has joined #zuul | 03:04 | |
| *** JasonCL has quit IRC | 03:09 | |
| *** JasonCL has joined #zuul | 03:19 | |
| *** JasonCL has quit IRC | 03:24 | |
| tristanC | fwiw, i just proposed a pre-release fedora package for zuulv3 here: https://src.fedoraproject.org/rpms/zuul/pull-request/1 | 03:33 |
| *** JasonCL has joined #zuul | 03:44 | |
| *** JasonCL has quit IRC | 03:49 | |
| *** JasonCL has joined #zuul | 04:01 | |
| *** JasonCL has quit IRC | 04:07 | |
| *** JasonCL has joined #zuul | 04:20 | |
| *** JasonCL has quit IRC | 04:24 | |
| *** JasonCL has joined #zuul | 04:44 | |
| *** JasonCL has quit IRC | 04:49 | |
| *** JasonCL has joined #zuul | 04:51 | |
| *** JasonCL has quit IRC | 04:58 | |
| *** JasonCL has joined #zuul | 05:48 | |
| *** JasonCL has quit IRC | 05:52 | |
| *** xinliang has quit IRC | 05:57 | |
| *** bhavik1 has joined #zuul | 06:07 | |
| *** xinliang has joined #zuul | 06:09 | |
| *** xinliang has quit IRC | 06:09 | |
| *** xinliang has joined #zuul | 06:09 | |
| *** JasonCL has joined #zuul | 06:12 | |
| *** JasonCL has quit IRC | 06:17 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Fix sql reporting with postgres https://review.openstack.org/532023 | 06:44 |
| *** JasonCL has joined #zuul | 06:46 | |
| *** JasonCL has quit IRC | 06:51 | |
| *** JasonCL has joined #zuul | 07:01 | |
| *** JasonCL has quit IRC | 07:06 | |
| *** jlk has quit IRC | 07:17 | |
| *** jlk has joined #zuul | 07:21 | |
| *** jlk has quit IRC | 07:21 | |
| *** jlk has joined #zuul | 07:21 | |
| *** JasonCL has joined #zuul | 07:41 | |
| gundalow | mordred: Could you please let me know a PR when it's created. I want to ensure it doesn't confuse Ansibullbot reading of CI status | 07:45 |
| gundalow | mattclay: fyi ^ | 07:46 |
| *** JasonCL has quit IRC | 07:46 | |
| *** hashar has joined #zuul | 07:50 | |
| *** JasonCL has joined #zuul | 08:36 | |
| *** JasonCL has quit IRC | 08:40 | |
| *** jpena|off is now known as jpena | 08:45 | |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul-jobs master: Add ansible-lint job https://review.openstack.org/532083 | 08:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul-jobs master: Add ansible-import-to-galaxy job https://review.openstack.org/532084 | 08:50 |
| openstackgerrit | Tristan Cacqueray proposed openstack-infra/zuul-jobs master: Add ansible-spec job https://review.openstack.org/532085 | 08:50 |
| rcarrillocruz | tristanC: heya, are you or fabien using the gitdriver to reconsume the upstream jobs yet ? | 09:08 |
| rcarrillocruz | i.e. avoiding the GH apps relationsships between zuuls | 09:08 |
| *** JasonCL has joined #zuul | 09:08 | |
| tristanC | rcarrillocruz: not yet, we still need a gerrit connection when deploying using openstack-infra/zuul-jobs | 09:08 |
| tristanC | rcarrillocruz: but because we also support offline deployment, we actually packaged and setup a known zuul-jobs locally by default | 09:09 |
| rcarrillocruz | ah, i see | 09:10 |
| *** JasonCL has joined #zuul | 09:10 | |
| tristanC | e.g. https://softwarefactory-project.io/r/gitweb?p=scl/zuul-jobs-distgit.git;a=tree | 09:12 |
| *** jlk has quit IRC | 09:18 | |
| *** jlk has joined #zuul | 09:26 | |
| *** jlk has quit IRC | 09:26 | |
| *** jlk has joined #zuul | 09:26 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: Refactor status functions, add web endpoints, allow params https://review.openstack.org/531390 | 09:34 |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: Add a separate module for node management commands https://review.openstack.org/531703 | 09:34 |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: [WIP] webapp: add optional admin endpoint https://review.openstack.org/531718 | 09:34 |
| *** bhavik1 has quit IRC | 09:47 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: zk: automatically retry command when connection is lost https://review.openstack.org/523640 | 09:54 |
| *** sshnaidm is now known as sshnaidm|afk | 10:07 | |
| *** fbo_ has quit IRC | 10:08 | |
| *** jkilpatr has quit IRC | 12:18 | |
| *** jpena is now known as jpena|lunch | 12:32 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: [WIP] webapp: add optional admin endpoint https://review.openstack.org/531718 | 12:35 |
| *** jkilpatr has joined #zuul | 12:54 | |
| *** sshnaidm|afk is now known as sshnaidm | 13:23 | |
| *** rlandy has joined #zuul | 13:31 | |
| *** jpena|lunch is now known as jpena | 13:32 | |
| *** fbo has joined #zuul | 13:33 | |
| *** dkranz has joined #zuul | 14:02 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: [WIP] webapp: add optional admin endpoint https://review.openstack.org/531718 | 14:18 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add error handling for pidfile https://review.openstack.org/532206 | 14:55 |
| pabelanger | leifmadsen_: mordred: corvus: ^something we noticed in testing docs, might be helpful to start adding some additional debug information to zuul users | 14:56 |
| *** sshnaidm is now known as sshnaidm|mtg | 15:01 | |
| *** rlandy is now known as rlandy|mtg | 15:05 | |
| *** JasonCL has quit IRC | 15:11 | |
| *** JasonCL has joined #zuul | 15:12 | |
| dmsimard | pabelanger: fyi https://review.openstack.org/#/c/532206/ is tangentially related to https://storyboard.openstack.org/#!/story/2001450 | 15:12 |
| dmsimard | https://storyboard.openstack.org/#!/story/2001445 as well, actually | 15:13 |
| pabelanger | dmsimard: we recently re-factored pidfiles in zuul, so that isn't the case anymore. But agree, we might need to update puppet-zuul or delete the old folders | 15:15 |
| pabelanger | everything is now in /var/run/zuul | 15:15 |
| tobiash | hmpf, sql reporter doesn't work with github | 15:16 |
| *** JasonCL has quit IRC | 15:16 | |
| tobiash | patchset is the git hash there and the db wants int :/ | 15:16 |
| pabelanger | also, I'm going to push up some systemd files for zuul today, which takes care of creating /var/run/zuul with systemd, but haven't tested across multiple systemd files | 15:17 |
| rcarrillocruz | erm, ok | 15:18 |
| dmsimard | upstart is so 2005 | 15:18 |
| dmsimard | 2006 according to wikipedia, I was close | 15:18 |
| rcarrillocruz | interesting tobiash , i was going to containerize mysql for zuul sql reporting in the next few days | 15:19 |
| *** smyers has quit IRC | 15:19 | |
| tobiash | looks like I need to look at how to write migration scripts | 15:21 |
| dmsimard | tobiash: why does the db want an int ? what do we have there for gerrit ? the change number ? | 15:22 |
| tobiash | dmsimard: gerrit has change number and patchset number | 15:22 |
| tobiash | where it made sense to use int and int | 15:23 |
| tobiash | github has not patchset number instead there the sha of the head of the pr is used | 15:23 |
| tobiash | which is no int ;) | 15:23 |
| dmsimard | tobiash: it would have the pr number? :/ | 15:23 |
| tobiash | pr number is like the change number | 15:24 |
| tobiash | ps is the change version aka the current head of the pr in gituhub | 15:24 |
| *** smyers has joined #zuul | 15:24 | |
| dmsimard | yeah | 15:24 |
| tobiash | so I need to change the patchset column to text | 15:24 |
| dmsimard | do we have the change-id somewhere in there ? | 15:25 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Use setup.py if we can't get the name from setup.cfg https://review.openstack.org/531936 | 15:25 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Capture and report errors in sibling installation https://review.openstack.org/532216 | 15:25 |
| dmsimard | at what point does the field just become a json blob? :D | 15:25 |
| tobiash | probably with visual source safe :-P | 15:25 |
| tobiash | that's the error btw: http://paste.openstack.org/show/641292/ | 15:26 |
| dmsimard | good idea to fix that before v3 is released | 15:27 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Capture and report errors in sibling installation https://review.openstack.org/532216 | 15:28 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Use setup.py if we can't get the name from setup.cfg https://review.openstack.org/531936 | 15:28 |
| corvus | tobiash: "alembic revision" will create a new migration file for you. you don't need to write a downgrade. here's an example of altering a column. you can copy gertty.dbsupport to get that function: http://git.openstack.org/cgit/openstack/gertty/tree/gertty/alembic/versions/56e48a4a064a_increase_status_field_width.py | 15:29 |
| tobiash | corvus: thanks :) | 15:29 |
| corvus | tobiash: and http://git.openstack.org/cgit/openstack/gertty/tree/gertty/dbsupport.py | 15:30 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Add support for a secure ZooKeeper configuration https://review.openstack.org/531882 | 15:38 |
| corvus | pabelanger: regarding 532206 i don't understand -- line 184 here should raise an exception if the pidfile is not writable, does it really not? http://git.openstack.org/cgit/openstack-infra/zuul/tree/zuul/cmd/__init__.py?h=feature%2Fzuulv3#n182 | 15:40 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Add error handling for pidfile https://review.openstack.org/532206 | 15:42 |
| pabelanger | corvus: yah, in testing, nothing happens for that logic. Not errors are raised | 15:44 |
| pabelanger | but I can look into why that is now | 15:44 |
| Shrews | corvus: pabelanger: yeah, i've seen that before too when testing the finger stuff. if the pid had the wrong owner, just a silent fail | 15:46 |
| * Shrews needs to make 1 more cup of coffee before a meeting | 15:50 | |
| corvus | Shrews, pabelanger: oh, because that should be entering the pidfile context, not daemon context. it should be "with pid". let's fix that first. ;) | 15:52 |
| pabelanger | cool, easy fixes are the best fixes | 15:54 |
| clarkb | Shrews: responded to your question on the secure file change | 15:58 |
| Shrews | clarkb: hrm, we can't really know which file the values come from. what if i put invalid values in the normal config that won't work? | 15:59 |
| clarkb | Shrews: ya that, I think you put invalid different values in the normal file, then override with good values in the secure file. Then just double check that you are using the good values | 15:59 |
| clarkb | in addition to things functioning | 15:59 |
| Shrews | clarkb: ok. i'll do that after this meeting | 16:01 |
| * Shrews reminds mordred and corvus about said meeting | 16:01 | |
| tobiash | corvus: is there a reason why there is no alembic.ini there? | 16:20 |
| *** jkilpatr_ has joined #zuul | 16:24 | |
| pabelanger | corvus: re: with pid, should we wrap in try / except or output default traceback? | 16:25 |
| *** jkilpatr has quit IRC | 16:26 | |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Change patchset column to string https://review.openstack.org/532237 | 16:32 |
| tobiash | corvus: ^^ I used alter_column as we use it already in another migration script | 16:33 |
| tobiash | If we want to support sqlite this should be a separate change | 16:33 |
| tobiash | however the dbsupport script didn't work out of the box due to the foreign key | 16:33 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Use setup.py if we can't get the name from setup.cfg https://review.openstack.org/531936 | 16:38 |
| *** rlandy|mtg is now known as rlandy | 16:41 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: [WIP] webapp: add optional admin endpoint https://review.openstack.org/531718 | 16:42 |
| clarkb | mordred: did you see my comment on ps1 of 531936? | 16:48 |
| mordred | clarkb: nope - looking now, thanks | 16:54 |
| *** jkilpatr has joined #zuul | 17:09 | |
| *** jkilpatr_ has quit IRC | 17:09 | |
| *** JasonCL has joined #zuul | 17:10 | |
| *** sshnaidm|mtg is now known as sshnaidm | 17:10 | |
| corvus | tobiash: i have no idea why there's no alembic.ini; seems like there probably should be one? | 17:11 |
| tobiash | it required at least a minimal one with the script location | 17:12 |
| corvus | tobiash: if you have one handy you can add, that's be great :) | 17:12 |
| tobiash | ok, no problem | 17:12 |
| openstackgerrit | Tobias Henkel proposed openstack-infra/zuul feature/zuulv3: Add alembic.ini https://review.openstack.org/532260 | 17:14 |
| tobiash | corvus: ^^ | 17:14 |
| *** hashar is now known as hasharAway | 17:17 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Fix error handling for pidfile https://review.openstack.org/532206 | 17:25 |
| pabelanger | corvus: Shrews: how's that^ | 17:25 |
| dmsimard | tobiash: I wonder if changing the column type has side effects (for example if the webapp exposes sqlreporter data and expects an int for some reason) | 17:26 |
| tobiash | dmsimard: just looked at the zuul-web code and it doesn't like like it's doing anything with this column | 17:29 |
| tobiash | so I believe it's save | 17:29 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Use setup.py if we can't get the name from setup.cfg https://review.openstack.org/531936 | 17:31 |
| corvus | pabelanger: why not just leave the bare exception? | 17:32 |
| corvus | pabelanger: or if you want to mask it, maybe something more helpful like "Unable to write to pidfile {pidfile} error: {e}" | 17:34 |
| pabelanger | corvus: http://paste.openstack.org/show/641397/ is would be the output, was a little verbose | 17:34 |
| pabelanger | but, can remove the print statement if you'd like | 17:35 |
| pabelanger | sure, let me try that | 17:35 |
| corvus | pabelanger: what's the output with your statement? | 17:35 |
| pabelanger | ERROR: Problem with pidfile (failed to create /var/run/zuul/zuul.pid) is current patchset | 17:35 |
| pabelanger | trying your updates now | 17:36 |
| corvus | pabelanger: that's the less useful of the two exceptions; if we reduce it to one, then we need the other one | 17:36 |
| corvus | pabelanger: basically, we should either let the exception through unaltered, or, if you want to mask it, the result should include "PermissionError: [Errno 13] Permission denied: '/var/run/zuul/zuul.pid' | 17:37 |
| corvus | that's the actionable part of the error | 17:37 |
| pabelanger | okay, lets pass though unaltered exception, let me update | 17:38 |
| dmsimard | we need something similar for /var/lib/zuul/* | 17:38 |
| dmsimard | documented it in https://storyboard.openstack.org/#!/story/2001445 | 17:39 |
| dmsimard | when I had remounted /var/lib/zuul into it's own partition on the reinstalled ze09, the permissions were not right and there was nothing in logs hinting at the issue | 17:40 |
| corvus | yep | 17:40 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Fix error handling for pidfile https://review.openstack.org/532206 | 17:40 |
| pabelanger | yah, I can get that one next | 17:40 |
| pabelanger | we had the same issue in quickstart docs | 17:41 |
| dmsimard | also, are those stack traces actually going to show up in logs ? | 17:41 |
| dmsimard | that's the important part | 17:41 |
| dmsimard | some traces appear when trying to start the process in foreground but they don't show up in logs .. | 17:41 |
| clarkb | dmsimard: the init scripts should set the permissions | 17:42 |
| dmsimard | maybe | 17:43 |
| clarkb | (so in addition to better error reporting, there is maybe a bug in the init script) | 17:43 |
| pabelanger | systemd has a setting now to automafically create /var/run directories | 17:43 |
| dmsimard | right, I was about to mention it's two separate improvement opportunities | 17:43 |
| clarkb | chown $USER $PIDFILE_DIR is in the script at least | 17:43 |
| clarkb | maybe it needs to set the group too? | 17:43 |
| dmsimard | clarkb: /var/lib/zuul != /var/run/zuul | 17:43 |
| pabelanger | but, currently is stdout | 17:43 |
| pabelanger | for pidfile | 17:43 |
| clarkb | oh the dir where the repos are kept, ya | 17:44 |
| clarkb | dmsimard: we should add that to the init script too | 17:44 |
| pabelanger | clarkb: zuul attempts to create it, but needs write permissions | 17:44 |
| pabelanger | so, we can raise exception if that fails | 17:45 |
| corvus | dmsimard: the pidfile check error will be printed to stderr before logging is set up, and will cause the process to exit nonzero. | 17:47 |
| pabelanger | ah, I think i see the issue. Currently daemon.DaemonContext will set stdout/stderr to None, which doesn't log some of our exceptions. | 17:54 |
| pabelanger | daemon.DaemonContext(pidfile=pid, stderr=sys.stderr) seems enable them again | 17:54 |
| pabelanger | then you get the following: http://paste.openstack.org/show/641416/ | 17:55 |
| clarkb | pabelanger: that is intentional because traditional daemonizing is supposed to close all file descriptors | 17:55 |
| clarkb | that assumption sort of changes when you start using tools like systemd though as they expect you will write logs to those fds and then handle them for you | 17:55 |
| clarkb | (upstart too) | 17:55 |
| pabelanger | clarkb: yah, for systemd it's not possible for it to daemonize itself and just use -d flag with zuul | 17:56 |
| clarkb | pabelanger: it should be possible to let systemd do the work and pair that with -d on the zuul size | 17:57 |
| SpamapS | upstart is dead | 17:57 |
| SpamapS | long live systemd | 17:57 |
| * SpamapS dies a little more inside | 17:57 | |
| clarkb | s/size/side/ | 17:57 |
| SpamapS | systemd _can_ handle traditional daemons, but they recommend either foreground or systemd-specific "readiness" | 17:57 |
| clarkb | basically what happened aiui is very few applications daemonized prperly so modern init systems said fine we'll do it for you | 17:57 |
| SpamapS | Yeah most things exit the parent before the children are actually "ready" | 17:58 |
| pabelanger | clarkb: yes, how I've been testing things myself with systemd and -d. So, unsure if we want to have the discussion of just letting systemd do that now | 17:58 |
| SpamapS | And many, as you say, don't do all the right things. | 17:58 |
| pabelanger | and remove daemon from zuul | 17:58 |
| corvus | nope | 17:58 |
| corvus | zuul will not depend on systemd | 17:58 |
| SpamapS | We should keep daemonization in zuul. How are we going to get people to run it on FreeBSD? ;-) | 17:58 |
| clarkb | ya I think we shouldn't remove that functionality, but we may not use it if using systemd unit files for zuul in those unit files | 17:59 |
| corvus | the current approach of detecting errors which prevent startup before daemonizing is the right one. | 17:59 |
| corvus | clarkb: that's a choice someone could make, though i don't think we should recommend it, at least, not yet. | 18:03 |
| clarkb | thats fine. I also think intentionally breaking daemonization (not closing fds like stderr) may not be the best approach if we aren't going to only support something like systemd | 18:06 |
| clarkb | but thats a tricky situation | 18:07 |
| corvus | clarkb: agreed. we should not break daemonization. | 18:07 |
| corvus | the approach should be to detect any errors which will prevent daemonization or logging further exceptions before performing the daemonization. | 18:08 |
| corvus | to that end, we may want to enter the pidfile context manager ourselves, before entering the daemonize context manager. so maybe something like "enter pidfile, verify log config, daemonize" | 18:10 |
| corvus | (ie, enter and stay entered, rather than enter/exit/enter) | 18:11 |
| *** jkilpatr has quit IRC | 18:17 | |
| *** jpena is now known as jpena|off | 18:29 | |
| mordred | corvus, clarkb, Shrews: woot - http://zuulv3.openstack.org/stream.html?uuid=8838e564afae4ef7adce227b6d42c147&logfile=console.log seems to be correctly running with src/github.com/ansible/ansible | 18:30 |
| *** jkilpatr has joined #zuul | 18:38 | |
| pabelanger | nice | 18:43 |
| clarkb | corvus: I think Icaptured the corrupted git repo situation and what we can do about it in zuul properly in https://storyboard.openstack.org/#!/story/2001453 | 18:47 |
| clarkb | is tox-py35-on-zuul still sometimes failing? | 18:48 |
| clarkb | mordred: is there already a story for logging more than command/shell? | 18:49 |
| clarkb | specific case on the zuulv3-issues ehterpad is the pat module | 18:49 |
| *** robled has quit IRC | 18:50 | |
| mordred | clarkb: you mean streaming the logs of non-command/shell? | 18:52 |
| mordred | clarkb: unfortunately no, not in zuul - we'll have to wait until we get streaming upstreamed into ansible to be able to support that | 18:52 |
| clarkb | ya | 18:53 |
| *** robled has joined #zuul | 18:55 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: [WIP] webapp: add optional admin endpoint https://review.openstack.org/531718 | 18:56 |
| clarkb | mordred: do we want a story to track that on the zuul side? | 18:56 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Use setup.py if we can't get the name from setup.cfg https://review.openstack.org/531936 | 19:01 |
| mordred | clarkb: yah - probably so | 19:01 |
| *** harlowja has joined #zuul | 19:33 | |
| *** harlowja_ has joined #zuul | 19:41 | |
| *** harlowja has quit IRC | 19:43 | |
| *** harlowja_ has quit IRC | 19:55 | |
| *** harlowja has joined #zuul | 19:55 | |
| mordred | corvus, clarkb, Shrews: http://logs.openstack.org/34/531934/3/check/zuul-tox-py35-ansible-devel/a1a9d08/tox/py35-siblings.txt <-- install-siblings correctly installed tip of ansible/ansible | 19:56 |
| Shrews | woo! | 19:56 |
| mordred | Shrews: now - the tests TOTALLY fail | 19:59 |
| mordred | butI think at the moment that's likely expected, given we had work to do to shift from 2.3 to 2.4 and whatnot | 20:00 |
| pabelanger | very cool | 20:02 |
| Shrews | mordred: oh, well that work is ready i think: https://review.openstack.org/531009 | 20:02 |
| corvus | if only we had cross-source-deps :) | 20:04 |
| pabelanger | could you set the branch to stable-2.3 for tox siblings? I haven't spent much time looking at sibligs | 20:04 |
| mordred | pabelanger: yup | 20:04 |
| pabelanger | good to know | 20:04 |
| mordred | corvus, pabelanger, Shrews, clarkb: how do we want to handle https://review.openstack.org/531009 and the corresponding upgrade of ansible on the executors to 2.4? | 20:05 |
| clarkb | mordred: probably put 2.4 on a single executor and see how it does? | 20:06 |
| clarkb | then rollout to the others if it is happy | 20:06 |
| pabelanger | I would imagine some playbooks some where would not react well to 2.4, so we should be on standby to help address that. | 20:07 |
| clarkb | ya ^ is probably the biggest concern | 20:08 |
| clarkb | maybe we need to make sure we expose the ansible version somewhere if not already? | 20:08 |
| clarkb | so that zuul users don't have to guess what version of ansible is breaking them | 20:08 |
| corvus | clarkb: i bet there's an ansible version variable already | 20:10 |
| corvus | http://docs.ansible.com/ansible/latest/playbooks_variables.html#ansible-version | 20:10 |
| corvus | maybe we should print that from our info role we run at the start of jobs? | 20:11 |
| mordred | yah | 20:11 |
| mordred | cause it sisn't in the host vars we print | 20:11 |
| Shrews | which is suprising | 20:12 |
| mordred | maybe in the job header ... since we don't actually need it per-test-node | 20:12 |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: Add Ansible version to job header https://review.openstack.org/532304 | 20:15 |
| mordred | somethign like that? | 20:15 |
| Shrews | lgtm | 20:16 |
| mordred | Shrews, clarkb, corvus, pabelanger, dmsimard: https://review.openstack.org/#/c/532216/ and https://review.openstack.org/#/c/531936 are good to go now | 20:17 |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: [WIP] webapp: add optional admin endpoint https://review.openstack.org/531718 | 20:34 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Add support for a secure ZooKeeper configuration https://review.openstack.org/531882 | 20:49 |
| Shrews | mordred: Wow. Your header change worked... but also didn't work | 20:54 |
| Shrews | that's neat | 20:55 |
| openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Switch to python3-dev / python3-devel in bindep.txt https://review.openstack.org/532310 | 20:57 |
| mordred | Shrews: oh - hah. nice | 20:58 |
| SpamapS | Does ansible-playbook have a "fail on warning" mode we could run jobs with at some point? Would be good to let people turn that on per-job to basically say "this job is warning free" that way it would be free from deprecated things. | 21:01 |
| mordred | SpamapS: not to my knowledge, no | 21:01 |
| SpamapS | Guess we could have a POST that fails after in log collection. | 21:02 |
| mordred | Shrews: weird ... http://logs.openstack.org/04/532304/1/check/openstack-infra-base-integration-centos-7/6081fe4/ara/result/fc028c65-a7ba-43ae-a221-adc2ec3ad023/ | 21:03 |
| SpamapS | don't know why that's a POST and not post-run ;) | 21:03 |
| mordred | SpamapS: you're putting your rests in your zuuls! | 21:03 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Add support for a secure ZooKeeper configuration https://review.openstack.org/531882 | 21:03 |
| mordred | dmsimard: I'm confused by the integration test errors for this change: https://review.openstack.org/#/c/532304/ | 21:05 |
| mordred | dmsimard: any chance they make sense to you? | 21:05 |
| dmsimard | that rings me a bell | 21:06 |
| dmsimard | something about include_vars including vars from the playbook directory instead of the role directory | 21:06 |
| SpamapS | mordred: or maybe I just like to yell POST | 21:06 |
| dmsimard | I've fixed something like that before, I forget where | 21:06 |
| dmsimard | trying to remember......... | 21:06 |
| clarkb | dmsimard: mordred it looks like fact gathering on localhost i afiling for trying to read things outside of the builds working dir? | 21:07 |
| dmsimard | well there you go -- http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/roles/configure-unbound/tasks/main.yaml#n41 | 21:07 |
| dmsimard | see how I prefix {{ role_path }} before.. | 21:07 |
| dmsimard | hmm | 21:07 |
| dmsimard | that's exactly where things are failing though | 21:08 |
| clarkb | oh ara claims its on the test node though | 21:08 |
| clarkb | and not on localhost | 21:08 |
| mordred | but include_vars is going to run on the executor | 21:08 |
| mordred | since it's an ansible include | 21:08 |
| dmsimard | is role_path not inside work/ ? | 21:09 |
| mordred | and I'm guessing role_path is including one of the source dirs perhaps? | 21:09 |
| dmsimard | it's wherever the role is located | 21:09 |
| mordred | dmsimard: it's a good question ... maybe we should print it | 21:09 |
| dmsimard | sure, let me add a debug somewhere in there | 21:09 |
| mordred | dmsimard: this is a fun one :) | 21:09 |
| dmsimard | mordred: the real issue is more sinister but more on that later | 21:10 |
| dmsimard | mordred: so the playbook that runs the tests is here: http://git.openstack.org/cgit/openstack-infra/openstack-zuul-jobs/tree/tests | 21:11 |
| dmsimard | and there's a vars dir there (because some tests have OS-specific things) | 21:11 |
| dmsimard | WELL, it turns out that when you do an include_vars, playbook/vars has a higher precedence than <role>/vars | 21:12 |
| mordred | \o/ | 21:12 |
| dmsimard | that was especially confusing, considering the include_vars is happening inside the role | 21:12 |
| dmsimard | I never managed to figure out if it was intended behavior | 21:12 |
| dmsimard | maybe a pro like Shrews or gundalow knows | 21:13 |
| mordred | dmsimard: I think it is - playbook local paths are supposed to win | 21:13 |
| mordred | dmsimard: we could cheat and just rename tests/vars to something else ... | 21:13 |
| dmsimard | but then if we do that, the include_vars from a playbook file would no longer work (or, well, we'd need to supply the different name there) | 21:14 |
| mordred | yah | 21:14 |
| dmsimard | IMO vars/ from the scope (playbook or role) should win, and then go seek out something else in order of priority | 21:14 |
| dmsimard | but I dunno | 21:14 |
| dmsimard | so anyway, role path print is here: https://review.openstack.org/#/c/532314/ | 21:15 |
| dmsimard | and also, I'd like to think that the current implementation works or else it wouldn't have merged so I'm not entirely sure why your thing is not working -- maybe the role_path is different because it's supplied by zuul (through zuul-jobs <-> openstack-zuul-jobs) while I would've landed the patch directly in o-z-j (in the work directory?) | 21:16 |
| Shrews | dmsimard: it's an easy mistake to assume that because I work for Ansible, that I am also an Ansible pro | 21:16 |
| mordred | dmsimard: how about I also add a role_path debug in my patch - we can compare them | 21:16 |
| dmsimard | mordred: sure, that's actually a good idea since my patch is in o-z-j | 21:17 |
| dmsimard | Shrews: I work on "computers" and I'm suddenly a pro in printers, faxes, VHS players and such :D | 21:17 |
| * dmsimard holidays are always painful | 21:18 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: DNM print role_path https://review.openstack.org/532319 | 21:18 |
| *** rlandy is now known as rlandy|brb | 21:35 | |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Move CRD tests to test_gerrit_legacy_crd https://review.openstack.org/531886 | 21:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Add skipped CRD tests https://review.openstack.org/531887 | 21:39 |
| openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Replace change.equals with __eq__ https://review.openstack.org/532333 | 21:39 |
| clarkb | Shrews: nodepool secure file lgtm | 21:43 |
| clarkb | Shrews: thanks | 21:43 |
| *** threestrands has joined #zuul | 21:44 | |
| *** threestrands has quit IRC | 21:44 | |
| *** threestrands has joined #zuul | 21:44 | |
| *** JasonCL has quit IRC | 21:49 | |
| mordred | clarkb, Shrews, corvus: any reason to not +A the secure.conf nodepool patch? it's got 3x+2 now | 21:51 |
| clarkb | mordred: I don't think so should handle no secure.conf just fine | 21:51 |
| mordred | ossum | 21:51 |
| clarkb | then we can add one whenever we add tls/auth to zk | 21:51 |
| mordred | dmsimard: role_path from my change: Role Path: /var/lib/zuul/builds/d1902e3d8aaa4390a4d7ffef408915bf/work/src/git.openstack.org/openstack-infra/zuul-jobs/roles/emit-job-header | 21:52 |
| dmsimard | and mine /var/lib/zuul/builds/24449f8a0fab45c98ec37631b61aa0e9/work/src/git.openstack.org/openstack-infra/openstack-zuul-jobs/roles/configure-unbound | 21:53 |
| * mordred is confused | 21:53 | |
| mordred | cause it says: Accessing files from outside the working dir /var/lib/zuul/builds/d1902e3d8aaa4390a4d7ffef408915bf/work is prohibited | 21:53 |
| mordred | but /var/lib/zuul/builds/d1902e3d8aaa4390a4d7ffef408915bf/work/src/git.openstack.org/openstack-infra/zuul-jobs/roles/emit-job-header is not outside of /var/lib/zuul/builds/d1902e3d8aaa4390a4d7ffef408915bf/work | 21:53 |
| dmsimard | mordred: your change doesn't expose what I talked about | 21:53 |
| dmsimard | mordred: we need to print role_path from within zuul-jobs | 21:53 |
| dmsimard | so your patch should depends-on my patch | 21:54 |
| mordred | kk. lemme do that | 21:54 |
| dmsimard | the behavior is around implied roles that zuul prepares I think | 21:54 |
| mordred | gotcha | 21:54 |
| SpamapS | everybody watch out. harlowja has fallen in love with YAML and Zuul | 21:55 |
| mordred | SpamapS: WOOT | 21:55 |
| harlowja | ummm | 21:55 |
| SpamapS | he's SO smooth at writing zuul jobs now | 21:55 |
| harlowja | i understand it, not sure love yet | 21:55 |
| harlowja | lol | 21:55 |
| * mordred hands harlowja a nice fluffy bunny | 21:55 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/zuul-jobs master: DNM print role_path https://review.openstack.org/532319 | 21:55 |
| harlowja | there are a few things that i think jenkins is doing better, that may be needed to really take it over | 21:55 |
| SpamapS | and even better, his zuul jobs write YAML to run kolla which is in YAML and it's all YAML | 21:55 |
| SpamapS | YAAAAMMMLLLL | 21:55 |
| harlowja | jenkins (though its plugins suck) is imho doing them better (ie, shared roles/libraries) | 21:56 |
| SpamapS | We need to get shirts that show the zuul monster eating yams. | 21:56 |
| harlowja | so that's one thing | 21:56 |
| SpamapS | We can call the zuul mascot "the yaminal" | 21:56 |
| harlowja | jenkins also has slightly different parallel abilities than zuul | 21:56 |
| *** JasonCL has joined #zuul | 21:56 | |
| harlowja | jenkins (in groovy) can dynamically grab slaves (of ?? any ?? amount) | 21:56 |
| harlowja | that's sorta different than nodeset | 21:56 |
| mordred | harlowja: by 'different' I think you mean 'falls over at a much lower scale' | 21:56 |
| SpamapS | harlowja: yeah that was one that I liked that you showed me. Having run-time nodesets basically. | 21:56 |
| harlowja | and what else, oh ya, the UI | 21:57 |
| harlowja | for those that want a nice UI, i think the blueocean stuff may be nicer (for simpler things) | 21:57 |
| harlowja | SpamapS ya, something like runtime nodesets | 21:57 |
| SpamapS | So instea dof like "I want a box named foo" it's "I want a box per value in this variable" | 21:57 |
| harlowja | https://jenkins.io/blog/2017/09/25/declarative-1/#parallel-stages | 21:57 |
| harlowja | ^ the groovy thing | 21:57 |
| harlowja | parallel really takes a map of functions (string key, value function) | 21:58 |
| SpamapS | would be good for DRY'ing when you have a generic job. | 21:58 |
| mordred | yah - so - this is a conceptual difference - there are things that jenkins treats as job functions in groovy that in zuul v3 almost always should actually be in the ansible rather than in the zuul | 21:58 |
| SpamapS | Also I do think there is a missing model in Zuul, and we've talked about it a few times, which is the "thing that follows a change from PR->package" | 21:59 |
| harlowja | mordred idk, need more description of what u think the difference is | 21:59 |
| harlowja | SpamapS PR -> package -> next -> next -> next | 21:59 |
| harlowja | https://jenkins.io/doc/pipeline/steps/ is also increasing (weekly) - for better/worse | 22:00 |
| mordred | harlowja: from my pov, jenkins pipelines are equivilent to ansible playbooks not to zuul jobs | 22:00 |
| harlowja | sorta like playbooks, yes | 22:00 |
| clarkb | SpamapS: jenkins doesn't have that model either though right? | 22:01 |
| *** rlandy|brb is now known as rlandy | 22:02 | |
| SpamapS | clarkb: build# | 22:02 |
| SpamapS | it does | 22:02 |
| mordred | yah, so basically, we have no interest in having zuul grow features in the way that jenkins pipelines have - that's ansible's job ... that said, the 'this job will want different types of nodes at different times' has come up from a few different people, so it seems lke something that needs to be addressed, even if addressing it is documentation or something | 22:02 |
| harlowja | the thing that jenkins doesn't have is the nodeset concept | 22:02 |
| harlowja | which is nice to be able to have | 22:02 |
| SpamapS | clarkb: because jenkins pipelines go across what zuul calls pipelines. | 22:02 |
| *** JasonCL has quit IRC | 22:02 | |
| harlowja | (in jenkins u can't at least not now request for a node then later use it, u have to use it immediatly) | 22:03 |
| harlowja | u can't get X nodes, then decide how u want to use them in a later part of the script (afaik) | 22:03 |
| clarkb | SpamapS: sort of, jenkins pipelines don't mix multiple triggers aiui its more a waterfall from some origination event right? | 22:03 |
| SpamapS | clarkb: yep, they take the change from start to stop | 22:03 |
| clarkb | SpamapS: oh so there are checkpoints where it waits for next event to happen? TIL | 22:04 |
| clarkb | harlowja: ya that limitation was why our multinode testing was always so weird | 22:04 |
| clarkb | harlowja: where we hacked in nodesets by giving jenkins the "primary" node then it could discover the rest of the set | 22:05 |
| harlowja | clarkb ya, i have a feeling they have to eventually fix that | 22:05 |
| SpamapS | clarkb: well more like, it does the thing and then keeps moving forward | 22:05 |
| harlowja | i mean, with groovy is so close, lol | 22:05 |
| harlowja | *it is so close | 22:05 |
| clarkb | SpamapS: except you don't want to build the package until the change merges, and you don't want to run th gate builds until the change has been approved | 22:05 |
| SpamapS | clarkb: so like, it gets a trigger to land a commit, builds stuff.. runs tests, lands commit, keeps going | 22:05 |
| clarkb | SpamapS: each of those requires a checkpoint to prevent the single event from cascading through it all | 22:06 |
| SpamapS | That's not how jenkins pipeline users use jenkins. :) | 22:06 |
| clarkb | SpamapS: right they are fundamentally different | 22:06 |
| mordred | yah. jenkins is an arbitrary ad-hoc automation platform with a web gui | 22:07 |
| clarkb | so jenkins pipelines don't cross zuul pipelines | 22:07 |
| clarkb | jenkins is just run jobs one after another and if that works for you thats great | 22:07 |
| SpamapS | Agreed, but the fact that I get to keep my gate artifacts as the end-artifacts is huge. | 22:07 |
| SpamapS | well they do cross from what we call gate to post. | 22:07 |
| SpamapS | the pipeline will gate and then post all in one pipeline run. | 22:08 |
| clarkb | I see | 22:08 |
| SpamapS | carrying the build# through | 22:08 |
| clarkb | so its a psecial case of checkpoint where gate has to pass to do post | 22:08 |
| clarkb | does jenkins do the git merging too? | 22:08 |
| corvus | the constant between a pre-merge pipeline and a post-merge pipeline is the change. one could store gate artifacts keyed by change and then use them in post. | 22:08 |
| SpamapS | that I don't know. | 22:08 |
| mordred | corvus: yah. that | 22:08 |
| corvus | SpamapS: is there something preventing that approach ^? | 22:08 |
| SpamapS | corvus: I don't have access to a consistent "change" in post. | 22:09 |
| SpamapS | I get a Branch effectively. | 22:09 |
| SpamapS | with a different sha | 22:09 |
| corvus | SpamapS: you use github, right? isn't there a pr-merged event or something? | 22:09 |
| SpamapS | no it's a push, IIRC | 22:09 |
| SpamapS | I could be wrong | 22:09 |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: webapp: add optional admin endpoint https://review.openstack.org/531718 | 22:10 |
| corvus | SpamapS: we created our post pipeline based on ref-updated due to a very specific gerrit behavior, i would prefer to build it on change-merged (and would probably spend some effort doing so if it were important to workflow) | 22:10 |
| clarkb | we could also do a lookup | 22:11 |
| clarkb | either in zuul or the job, but find out what the change was for this project + branch + changeid combo | 22:11 |
| clarkb | since that is the key tuple for looking up changes | 22:11 |
| corvus | clarkb: yeah, that's a possibility | 22:12 |
| corvus | there is a pull request closed event, however i'm not sure of 2 things: 1) what would be involved in differentiating merged vs abandoned. 2) whether the merge sha is provided (this is the same piece of info missing from gerrit that led us to use ref-updated) | 22:13 |
| SpamapS | Yeah IIRC we punted on that for Bonny for similar reasons. | 22:14 |
| corvus | at any rate, generally speaking, having a trigger with both the change id and sha solves a lot of problems -- it makes the workflow under discussion easier, and it lets us do things like provide retrospective reports on changes. | 22:14 |
| corvus | zuul's drivers are getting sophisticated enough it may be worth correcting the github and gerrit deficiencies in this by doing lookups. | 22:15 |
| SpamapS | that would be nice. | 22:15 |
| SpamapS | for now, just ignore harlowja | 22:15 |
| corvus | 3.1 :) | 22:15 |
| SpamapS | he's just jelly ;) | 22:15 |
| mordred | SpamapS: I've always thought of him more as a marmalade myself | 22:16 |
| *** dkranz has quit IRC | 22:19 | |
| harlowja | ya, ignore me, lol | 22:34 |
| SpamapS | you see the burning envy on display^^ | 22:34 |
| SpamapS | ;) | 22:34 |
| *** JasonCL has joined #zuul | 22:43 | |
| openstackgerrit | Matthieu Huin proposed openstack-infra/nodepool feature/zuulv3: webapp: add optional admin endpoint https://review.openstack.org/531718 | 22:46 |
| clarkb | mordred: fyi https://storyboard.openstack.org/#!/story/2001459 filed for the apt et al task logging | 22:57 |
| *** jlk is now known as Guest76 | 23:03 | |
| *** hasharAway has quit IRC | 23:04 | |
| Guest76 | corvus: Sorry I missed it. I'm at orientation for my new gig all week. | 23:05 |
| *** Guest76 is now known as jlk | 23:06 | |
| jlk | uh | 23:06 |
| jlk | corvus: Sorry I missed, I'm at orientation for my new gig all this week. | 23:06 |
| clarkb | Shrews: that secure.conf change retry limited on a pep8 change? | 23:15 |
| clarkb | current opnstack gate also has a timeout on a unittest change | 23:15 |
| clarkb | we may need to dig in and look at zuul stability | 23:15 |
| clarkb | load average: 25.27, 17.15, 12.63 on ze10 may explain http://logs.openstack.org/36/531836/1/gate/cross-neutron-py27/605a6da/job-output.txt.gz timing out | 23:18 |
| clarkb | hrm actually that job may just have a timeout that is too short to run neutron unittest but the ze10 load does seem high | 23:20 |
| *** jlk is now known as Guest76 | 23:29 | |
| mordred | dmsimard: why did it work this time????? https://review.openstack.org/#/c/532319/ | 23:40 |
| mordred | dmsimard: like, there is no reason why depends-on on your ozj job should fix it ... | 23:46 |
| mordred | dmsimard: OH ... | 23:46 |
| SpamapS | hm... trying to add sql reporter to my zuul and it is kind of just freezing up | 23:46 |
| mordred | SpamapS: also, be aware there is a github/sql-reporter bug that tobiash was working on a fix for | 23:46 |
| mordred | SpamapS: https://review.openstack.org/#/c/532237/ | 23:47 |
| mordred | SpamapS: column type issue | 23:47 |
| SpamapS | aha | 23:47 |
| SpamapS | well then I will just back that change right out | 23:47 |
| SpamapS | well actually | 23:48 |
| * SpamapS fires up gertty | 23:48 | |
| mordred | dmsimard: having ozj in the depends-on path means that zuul will have checked out the role potentially into a different path | 23:48 |
| mordred | SpamapS: I just now hit +A - I think you want to *add* that change to your local install if you want to use sql-reporter and github | 23:48 |
| mordred | SpamapS: I don't know why it would cause a hang though | 23:49 |
| mordred | but I didn't do any of the troubleshooting on it | 23:49 |
| SpamapS | Yeah I don't think that's the hang problem | 23:49 |
| SpamapS | I'm stracing to figure that out | 23:49 |
| SpamapS | because there's no logs or anything | 23:49 |
| SpamapS | it's just stuck in 'futex' | 23:49 |
| *** ianw has quit IRC | 23:50 | |
| mordred | SpamapS: silly futex | 23:51 |
| SpamapS | oh my.. | 23:51 |
| SpamapS | something is eating my stderr | 23:51 |
| SpamapS | [pid 7915] write(2, "ImportError: No module named 'py"..., 39ImportError: No module named 'pymysql' | 23:51 |
| *** ianw has joined #zuul | 23:51 | |
| mordred | SpamapS: maybe the stderr eating is related to the earlier conversation about appropriate behaviors in daemonizing and non-daemonizing modes? (I will admit I wasn't paying perfect attention to them) | 23:52 |
| SpamapS | mordred: I'm running under systemd with -d | 23:53 |
| SpamapS | and nor was I paying attention | 23:53 |
| SpamapS | for that matter | 23:53 |
| SpamapS | I forgot to pay attention to my stomach | 23:53 |
| * SpamapS afks for refueling | 23:53 | |
| openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Change patchset column to string https://review.openstack.org/532237 | 23:57 |
| SpamapS | \o/ | 23:57 |
| * SpamapS rebases local branch on that | 23:57 | |
| openstackgerrit | Clint 'SpamapS' Byrum proposed openstack-infra/zuul feature/zuulv3: Slack driver https://review.openstack.org/530521 | 23:59 |
| openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Add support for a secure ZooKeeper configuration https://review.openstack.org/531882 | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!