openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 00:25 |
---|---|---|
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 00:31 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 00:40 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher! https://review.openstack.org/437700 | 00:49 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 00:53 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 00:56 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 00:59 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 01:02 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 01:09 |
pabelanger | okay, that is it for tonight. Log collection from worker to zuul-launcher is working | 01:18 |
SpamapS | sweet | 01:19 |
openstackgerrit | K Jonathan Harker proposed openstack-infra/zuul feature/zuulv3: Add support for 'repos' job configuration https://review.openstack.org/436712 | 01:28 |
openstackgerrit | K Jonathan Harker proposed openstack-infra/zuul feature/zuulv3: A couple small test improvements https://review.openstack.org/437592 | 01:59 |
mordred | pabelanger: you are amazing | 02:11 |
openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Change log streaming link to finger protocol https://review.openstack.org/437764 | 02:46 |
openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: WIP Put real information in log streaming link https://review.openstack.org/437765 | 02:46 |
mordred | jeblair: ^^ am I on the right track with change #2 there? | 02:46 |
mordred | jlk: oh - piddle - also, I think I changed too many tests in the first change | 02:47 |
mordred | gah | 02:47 |
mordred | jeblair: ^^ | 02:48 |
jeblair | mordred: yes i think so | 02:55 |
mordred | awesome | 02:55 |
jeblair | mordred: can probably drop the trailing /0 | 02:56 |
mordred | what is the trailing /0 ? | 02:56 |
jeblair | mordred: also, i reckon we probably want finger://unique@server -- job isn't really necessary there | 02:57 |
mordred | jeblair: ++ | 02:57 |
jeblair | mordred: 'finger://{unique}@{server}/job/{job}/0/' --> 'finger://{unique}@{server}' | 02:58 |
mordred | ++ | 02:59 |
mordred | jeblair: how do I get the list of builds from a job? | 03:05 |
mordred | or, rather - let me ask that differenting | 03:06 |
openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Change log streaming link to finger protocol https://review.openstack.org/437764 | 03:08 |
mordred | jeblair: I put in a HELP comment in that one - I'll keep poking around but if you hav ea quick/easy answer that would be neat | 03:08 |
jeblair | mordred: status_jobs[X]['uuid'] should have it | 03:14 |
jeblair | mordred: that's *barely* testing that, but it's something :) | 03:14 |
jeblair | mordred: if we were to do something more robust, it would involve also comparing that to the internal data structure and is a bit more work... considering we didn't think it was worthwhile to do that in the rest of the test, checking it against the sibling json data is probably fine :) | 03:15 |
mordred | jeblair: :) cool, thanks! (I was thinking about checking the internal data structure, but happy to keep it simple for the first pass) | 03:21 |
jeblair | mordred: ya, if we wanted to do that, it's self.sched.abide[tenant].layout.pipelines... | 03:22 |
openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Change log streaming link to finger protocol https://review.openstack.org/437764 | 03:24 |
*** saneax-_-|AFK is now known as saneax | 04:13 | |
*** saneax is now known as saneax-_-|AFK | 04:25 | |
*** hashar has joined #zuul | 08:45 | |
*** Cibo_ has joined #zuul | 09:04 | |
*** saneax-_-|AFK is now known as saneax | 09:07 | |
*** Cibo_ has quit IRC | 09:09 | |
*** pleia2 has quit IRC | 09:27 | |
*** anteaya has quit IRC | 09:27 | |
*** pleia2 has joined #zuul | 09:28 | |
*** anteaya has joined #zuul | 09:30 | |
*** Cibo_ has joined #zuul | 11:07 | |
*** openstackgerrit has quit IRC | 12:18 | |
rcarrillocruz | clarkb: https://review.openstack.org/#/c/435933/ is slowly getting there | 12:19 |
rcarrillocruz | i'll clean up a bit the task names and other stuff | 12:19 |
rcarrillocruz | but is now passing the overlay setup and running the full tempest | 12:19 |
rcarrillocruz | (on dvr) | 12:19 |
rcarrillocruz | i'll ping you again soon when i think i've ironed out a few cosmetic things, but let me know if you spot any functional issue | 12:20 |
clarkb | ok | 12:40 |
*** Cibo_ has quit IRC | 13:11 | |
pabelanger | morning | 13:17 |
pabelanger | down on level 1 ATM | 13:17 |
Shrews | looks like lots of good things happened yesterday | 13:25 |
*** openstackgerrit has joined #zuul | 13:39 | |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Create python-linters job https://review.openstack.org/437059 | 13:39 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Copy logs to launcher https://review.openstack.org/437700 | 13:39 |
jeblair | Shrews: yes, and more today i hope! | 13:50 |
jeblair | pabelanger: sorry, i'm getting a late start; i'll try to join you in about an hour | 13:51 |
pabelanger | jeblair: no rush | 13:51 |
*** saneax is now known as saneax-_-|AFK | 14:18 | |
openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Delete node request locks after a delay https://review.openstack.org/437922 | 14:34 |
Shrews | jeblair: pabelanger: I realized last night that we could hit the scenario detailed in the commit message of ^^^^ | 14:35 |
openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Delete node request locks after a delay https://review.openstack.org/437922 | 14:38 |
openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool feature/zuulv3: Delete node request locks after a delay https://review.openstack.org/437922 | 14:39 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 15:03 |
*** jamielennox|away is now known as jamielennox | 15:07 | |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 15:07 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 15:25 |
jeblair | Shrews: remind me again why we need to cleanup request locks out of band? i get that since a request can disappear and the lock isn't a child, it needs to be separately deleted; but if we locked the request, i know that we also check to make sure it's still around. wouldn't we unlock it in the normal course of events? | 15:26 |
Shrews | jeblair: unlock != deleted | 15:27 |
jeblair | Shrews: should we delete the lock file at the time we unlock? | 15:27 |
Shrews | jeblair: the lock is a directory (non-ephemeral) with ephemeral children | 15:27 |
jeblair | Shrews: ah, that could be bad because it may race a new lock, yeah? | 15:27 |
Shrews | yep | 15:28 |
jeblair | okay, i think i'm caught up. thanks. :) | 15:28 |
openstackgerrit | Merged openstack-infra/nodepool feature/zuulv3: Delete node request locks after a delay https://review.openstack.org/437922 | 15:32 |
Shrews | oh, pabelanger... you were asking about test openshift accounts. i learned there is this thing: https://github.com/minishift/minishift | 15:35 |
jeblair | mordred: pabelanger suggests that we publish zuulv3 logs to an apache running on zuulv3-dev for a while so we don't put the static.o.o key on the system until we're more secure | 15:37 |
mordred | jeblair: wfm | 15:37 |
jeblair | mordred: clarkb has questions about bubblewrap; specifically, the permissions needed to run it. | 15:40 |
mordred | jeblair: so - with the finger change - we seem to be reporting streaming link and final publication link the same | 15:41 |
jeblair | mordred: yeah, we need to do something about the log path pattern | 15:41 |
clarkb | right if its setuid and escalating via that binary isn't the exposure surface still quite large? | 15:41 |
mordred | jeblair, clarkb: it should not require elevated permissions to run | 15:41 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 15:42 |
mordred | clarkb: well - I think the idea is that it's a small setuid binary that is easily auditable | 15:42 |
*** Cibo_ has joined #zuul | 15:42 | |
clarkb | ok I think where I am mostly confused is this is how docker operates too iirc. And docker lets you do all the things | 15:44 |
clarkb | but in this case the assertion is bubblewrap restricts those operations to some minimal set that is harder t oabuse and is small and audited to help avoid break outs? | 15:45 |
clarkb | the other thing I noticed is it uses tmpfs which will restrict the number of concurrent launchers being run within bubblewrap (based on available memory and size of job storage) | 15:46 |
*** Cibo_ has quit IRC | 15:47 | |
mordred | clarkb: so we may want to make a director for the workdir and have it mount that in | 15:47 |
mordred | but yeah - restricting the operations is key - and also it's trying to make use of user_capabilities as the kernel picks them up | 15:47 |
mordred | so over time it should trend towards using the non-privileged containment features of the kernel | 15:48 |
mordred | clarkb: there's a guy on my team at RH who has worked on it some I could ping if we want to go deeper with questions - and also tristanC landed a patch to it recently, so he may also know things | 15:49 |
clarkb | ya jeblair asked if tristanC was here (is not unfortunately) | 15:49 |
*** Cibo_ has joined #zuul | 15:49 | |
clarkb | mordred: I think it would be good to properly understand what the risk is there before we assert it fixes some of the problems we face | 15:49 |
mordred | jeblair: so for now should we just update the tests of the success url to use the same status url and then deal with success url later? | 15:51 |
mordred | clarkb: agree | 15:51 |
mordred | clarkb: I pinged vbatts and will try to grab him so we can ask some deeper questions | 15:53 |
*** yolanda has quit IRC | 15:53 | |
rbergeron | random squirrel thought: (maybe not since you guys seem to be talking about kubernetes things?) -- i saw this thing yesterday https://github.com/coreos/zetcd | 15:57 |
rbergeron | "A ZooKeeper "personality" for etcd. Point a ZooKeeper client at zetcd to dispatch the operations on an etcd cluster. | 15:58 |
rbergeron | " | 15:58 |
mordred | rbergeron: wow. that seems like a TERRIBLE idea | 15:59 |
rbergeron | not sure if that might help with kubernetes cluster things (as in, hook zookeeper to etcd so we don't have to have multiple things?) but um, who knows? | 15:59 |
rbergeron | or it could be a terrible idea? | 15:59 |
rbergeron | :) | 15:59 |
rbergeron | i just bring things to y'all, sometimes it's terrible, that happens :) | 15:59 |
mordred | rbergeron: well, the reason we chose zookeeper instead of etcd is that zookeeper can support fair locking and etcd can't do so fully safely, so emulating zookeeper with an etcd backend would work for many things, but then the semantics of operation of some of the things would fail in dangerous ways | 16:00 |
mordred | AIUI | 16:00 |
clarkb | ya there are lock queuing and thundering herd issues | 16:00 |
mordred | clarkb: chatted briefly with vbatts, he suggested we grab alexlarsson on Monday (he's in Sweden so timezones) | 16:02 |
mordred | also apparently it will function without the setuid, but may not be able to do all the things | 16:02 |
clarkb | interesting | 16:02 |
mordred | I think it uses the setuid power to do the things that haven't grown user-capability support yet | 16:03 |
clarkb | ya so feature set likely dependent on kernel version | 16:03 |
mordred | yah | 16:03 |
clarkb | mordred: you know apt and dib things can you review https://www.openstack.org/software/ocata/ ? | 16:05 |
clarkb | dangit | 16:05 |
clarkb | https://review.openstack.org/#/c/437425/1 | 16:05 |
clarkb | is the actual url | 16:05 |
mordred | clarkb: I don't know what /var/lib/apt/list is | 16:06 |
clarkb | mordred: its the indexes from repos | 16:06 |
mordred | mordred@camelot:~/src/github.com/ansible/ansible$ ls /var/lib/apt/list | 16:06 |
mordred | ls: cannot access '/var/lib/apt/list': No such file or directory | 16:06 |
mordred | not on my computer it isn't | 16:06 |
clarkb | neat | 16:07 |
mordred | :) | 16:07 |
clarkb | do you have a /var/lib/apt? | 16:07 |
mordred | yes. AH - it's lists | 16:07 |
mordred | and yes, there they are | 16:07 |
openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Change log streaming link to finger protocol https://review.openstack.org/437764 | 16:10 |
jeblair | mordred: jhesketh is going to work on the reporting url | 16:18 |
mordred | jeblair: \o/ | 16:25 |
jeblair | mordred: https://review.openstack.org/437707 | 16:30 |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add --safe-links to synchronize invocations https://review.openstack.org/437568 | 16:33 |
*** bhavik1 has joined #zuul | 16:35 | |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add support for 'repos' job configuration https://review.openstack.org/436712 | 16:36 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Fix synchronize action issue with mode https://review.openstack.org/437707 | 16:38 |
*** abregman has joined #zuul | 16:51 | |
openstackgerrit | Monty Taylor proposed openstack-infra/zuul feature/zuulv3: Change log streaming link to finger protocol https://review.openstack.org/437764 | 16:55 |
mordred | jeblair: ok. it's going to pass tests this time | 16:55 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Add more variables to ansible https://review.openstack.org/437999 | 17:02 |
* jeblair lunches | 17:03 | |
*** abregman is now known as abregman|afk | 17:05 | |
*** bhavik1 has quit IRC | 17:06 | |
mordred | when folks get back, they may or may not enjoy these: https://review.openstack.org/#/c/436455/ and https://review.openstack.org/#/c/436452/ | 17:33 |
*** yolanda has joined #zuul | 17:40 | |
*** Cibo_ has quit IRC | 17:47 | |
*** hashar has quit IRC | 17:48 | |
*** yolanda has quit IRC | 17:52 | |
*** yolanda has joined #zuul | 17:54 | |
jeblair | mordred: how might we go about having a playbook provide structured data back to the launcher (eg, the logfile copy playbook returns a url back to the launcher)? | 18:07 |
jeblair | mordred: also, jenkinsectomy++ :) | 18:07 |
jeblair | mordred: i think i might suggest we hold off on the infra-manual change for a bit longer though since currently "Jenkins" does report, and it may be confusing to new users. | 18:09 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: WIP testing log urls https://review.openstack.org/438028 | 18:10 |
jhesketh | jeblair: https://review.openstack.org/#/c/438028/1 | 18:12 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: WIP testing log urls https://review.openstack.org/438028 | 18:23 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 18:26 |
pabelanger | jhesketh: jeblair: can we land 437707 please | 18:29 |
jeblair | mordred: wow. what is going on in the unit test output of 437764? | 18:29 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path! https://review.openstack.org/437938 | 18:33 |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix synchronize action issue with mode https://review.openstack.org/437707 | 18:36 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: WIP testing log urls https://review.openstack.org/438028 | 18:38 |
SpamapS | jeblair: timed out by zuul? | 18:41 |
openstackgerrit | James E. Blair proposed openstack-infra/zuul feature/zuulv3: Fix keyerror with synchronize https://review.openstack.org/438043 | 18:41 |
jeblair | SpamapS: ^ probably | 18:41 |
SpamapS | oh, yeah that makes sense | 18:42 |
mordred | jeblair: that's impressive | 18:45 |
*** yolanda has quit IRC | 18:47 | |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 18:47 |
mordred | jeblair: as for the other thing, there 2 options I can think of are a) we ignore stdout from our callback plugin anyway - have it write json to stdout that we can then consume b) have the logfile copy playbook write a file with info into a known location that zuul can read from | 18:49 |
mordred | jeblair: neither are _great_ options though - let me noodle on it for a bit ... | 18:49 |
* mordred afks for an errand | 18:49 | |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path! https://review.openstack.org/437938 | 18:49 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path https://review.openstack.org/437938 | 18:54 |
pabelanger | 2017-02-24 18:55:05,170 DEBUG zuul.AnsibleJob: Ansible output: [WARNING]: Failure using method (v2_playbook_on_task_start) in callback plugin | 18:55 |
pabelanger | mordred: ^ | 18:55 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path. https://review.openstack.org/437938 | 18:58 |
SpamapS | mordred: fyi, bsd finger reads and writes one character at a time, until the finger server closes the socket | 18:59 |
SpamapS | while ((c = getc(fp)) != EOF) { | 19:00 |
SpamapS | mordred: so we probably don't need a rust one that does that more efficiently ;) | 19:00 |
openstackgerrit | K Jonathan Harker proposed openstack-infra/zuul feature/zuulv3: A couple small test improvements https://review.openstack.org/437592 | 19:01 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: WIP testing log urls https://review.openstack.org/438028 | 19:02 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path. https://review.openstack.org/437938 | 19:03 |
*** abregman|afk has quit IRC | 19:05 | |
*** yolanda has joined #zuul | 19:06 | |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path. https://review.openstack.org/437938 | 19:07 |
*** yolanda has quit IRC | 19:12 | |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path. https://review.openstack.org/437938 | 19:12 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path. https://review.openstack.org/437938 | 19:15 |
openstackgerrit | Paul Belanger proposed openstack-infra/zuul feature/zuulv3: Test rsync_path. https://review.openstack.org/437938 | 19:20 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Simplify the log url https://review.openstack.org/438028 | 19:33 |
pabelanger | http://zuulv3-dev.openstack.org/logs/ | 19:33 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Simplify the log url https://review.openstack.org/438028 | 19:34 |
openstackgerrit | Joshua Hesketh proposed openstack-infra/zuul feature/zuulv3: Simplify the log url https://review.openstack.org/438028 | 19:38 |
*** herlo has joined #zuul | 19:52 | |
*** herlo has joined #zuul | 19:52 | |
*** herlo has left #zuul | 19:52 | |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix keyerror with synchronize https://review.openstack.org/438043 | 19:55 |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Add more variables to ansible https://review.openstack.org/437999 | 19:56 |
*** jamielennox is now known as jamielennox|away | 19:57 | |
* jeblair is at airpoirt | 20:19 | |
jesusaur | pabelanger: cool! | 20:23 |
openstackgerrit | Merged openstack-infra/zuul feature/zuulv3: Fix getInfoRefs call from getProjectBranches https://review.openstack.org/436801 | 20:30 |
*** eventingmonkey has joined #zuul | 20:57 | |
dmsimard | pabelanger: nice, I gotta feed ara in there now :p | 22:44 |
openstackgerrit | K Jonathan Harker proposed openstack-infra/zuul master: Optionally require a zuul-ref when cloning https://review.openstack.org/227955 | 23:12 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!