| *** jamielennox is now known as jamielennox|away | 00:07 | |
| *** harlowja has quit IRC | 00:09 | |
| *** jamielennox|away is now known as jamielennox | 00:11 | |
| *** jasondotstar has quit IRC | 00:11 | |
| *** zaro has quit IRC | 00:12 | |
| *** zaro has joined #zuul | 00:12 | |
| *** _ari_ has quit IRC | 00:12 | |
| *** _ari_ has joined #zuul | 00:17 | |
| *** jasondotstar has joined #zuul | 00:18 | |
| *** harlowja has joined #zuul | 00:25 | |
| *** saneax is now known as saneax-_-|AFK | 01:48 | |
| *** Shuo has quit IRC | 02:44 | |
| *** Cibo has quit IRC | 05:15 | |
| *** Cibo has joined #zuul | 05:28 | |
| *** bhavik1 has joined #zuul | 05:54 | |
| *** bhavik1 has quit IRC | 06:03 | |
| *** saneax-_-|AFK is now known as saneax | 06:45 | |
| *** hogepodge has quit IRC | 07:56 | |
| *** Cibo_ has joined #zuul | 08:16 | |
| *** bhavik1 has joined #zuul | 09:06 | |
| *** bhavik1 has quit IRC | 09:23 | |
| *** hashar has joined #zuul | 10:26 | |
| openstackgerrit | Merged openstack-infra/zuul: Add Zookeeper to tests https://review.openstack.org/413241 | 11:55 |
|---|---|---|
| *** saneax is now known as saneax-_-|AFK | 12:45 | |
| *** jamielennox is now known as jamielennox|away | 14:14 | |
| *** saneax-_-|AFK is now known as saneax | 14:27 | |
| *** Shuo has joined #zuul | 16:39 | |
| *** dmsimard is now known as dmsimard|afk | 17:08 | |
| openstackgerrit | James E. Blair proposed openstack-infra/nodepool: Merge branch 'master' into feature/zuulv3 https://review.openstack.org/419054 | 17:14 |
| jeblair | Shrews: check that out locally and see if the tree looks correct please | 17:14 |
| jeblair | here's the list of files touched: http://paste.openstack.org/show/594598/ | 17:17 |
| Shrews | jeblair: quick inspection looks good | 17:21 |
| Shrews | gerrit doesn't make it easy to compare, but some 'show-branch' comparisons show everything to be there | 17:23 |
| jeblair | Shrews: gertty makes it easy | 17:24 |
| jeblair | Shrews: it will conflict with your current stack. do you want to land the merge commit and rebase the stack, or land the stack and i can re-do the merge commit? | 17:25 |
| Shrews | jeblair: i'll keep it easy on you and just rebase my stuff | 17:26 |
| jeblair | cool -- the merge commit didn't have any conflicts, so i kind of lean toward that anyway -- it'll be easier and safer to resolve the conflicts in the individual commits in your stack | 17:26 |
| jeblair | clarkb: would you mind approving 419054? | 17:27 |
| openstackgerrit | Merged openstack-infra/nodepool: Update configuration docs for network and pool settings https://review.openstack.org/417691 | 17:30 |
| clarkb | jeblair: sure just git back in from dealing with snow | 17:38 |
| Shuo | Is there a talk session, either in OpenStack Summit or outside, that explains the details of the current OpenStack Infra? (under the hood architecture details, daily operational statistics, and etc.) | 17:40 |
| Shuo | that I can find on Youtube? | 17:40 |
| *** hashar has quit IRC | 17:41 | |
| clarkb | pleia2: ^ do you have a recent one handy? | 17:42 |
| * Zara finds https://www.youtube.com/watch?v=b5PRk69EMuA from Barcelona | 17:42 | |
| Zara | (pleia2's keynote on the infra) | 17:43 |
| clarkb | oh right that | 17:45 |
| clarkb | jeblair: I am impressed that with that many changes the only resolved conflict was the branch in .gitreview | 17:47 |
| jeblair | yeah, i had it easy :) | 17:49 |
| clarkb | (I think that means we are properly compartmentalizing the work post initial merging back and forth) | 17:49 |
| *** Cibo_ has quit IRC | 17:58 | |
| *** saneax is now known as saneax-_-|AFK | 18:03 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool: Drop 'template' from default value of template-hostname https://review.openstack.org/411515 | 18:07 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool: Drop 'template' from default value of template-hostname https://review.openstack.org/411515 | 18:08 |
| pabelanger | jeblair: clarkb: mind reviewing https://review.openstack.org/#/c/414385/ ? Drops our RAM usage for ubuntu DIBs in nodepool dsvm, and fixes a bug in our check script | 18:09 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool: Support AFS mirrors for nodepool diskimages https://review.openstack.org/414273 | 18:19 |
| clarkb | pabelanger: rereading the help on nova flavor-create any idea what root disk size 0 means? | 18:24 |
| clarkb | pabelanger: maybe we get a disk that grows as necessary? | 18:24 |
| clarkb | (its what the old flavor did so approving change now, but curious what that is doing) | 18:24 |
| pabelanger | clarkb: not sure actually, it was just copied from the previous flavor-create command. | 18:26 |
| pabelanger | will read up and find out | 18:26 |
| clarkb | pabelanger: ya its not a regression so not worried just confused :) | 18:26 |
| openstackgerrit | Merged openstack-infra/nodepool: Reduce min-ram for ubuntu images for devstack https://review.openstack.org/414385 | 18:27 |
| openstackgerrit | Merged openstack-infra/nodepool: Merge branch 'master' into feature/zuulv3 https://review.openstack.org/419054 | 18:30 |
| Shuo | Zara: thanks | 18:52 |
| Shrews | stack rebase bomb inbound.... | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Remove --no-builder nodepoold option https://review.openstack.org/418013 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add ZK API methods for node requests https://review.openstack.org/416671 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Handle provider removal from config https://review.openstack.org/418112 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Create per-provider ZK work threads https://review.openstack.org/417994 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Remove Gearman and ZMQ https://review.openstack.org/417128 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Set valid states per data model class https://review.openstack.org/417482 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Disable tests for nodepoold changes https://review.openstack.org/417108 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Remove --no-launches nodepoold option https://review.openstack.org/418001 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Minor code cleanup https://review.openstack.org/417199 | 18:53 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add framework for handling node requests https://review.openstack.org/418585 | 18:53 |
| Shrews | *BOOM* | 18:54 |
| mordred | Shrews: nice | 18:54 |
| Shrews | mordred: it really is quite mordred-esque, isn't it? | 18:54 |
| mordred | Shrews: if it was mordred, you'd have messed up pep8 on the first patch and follow that bomb with an immediate second one :) | 18:55 |
| Shrews | there's still a chance | 18:58 |
| *** hashar has joined #zuul | 19:52 | |
| *** morgan_ is now known as morgan | 19:58 | |
| Shuo | I found (and watched) this talk very helpful, which help me answer some preliminary question and think deeper ones, just to answer my own previous question | 20:56 |
| *** dmsimard|afk is now known as dmsimard | 21:00 | |
| *** jamielennox|away is now known as jamielennox | 21:01 | |
| Shuo | I found (and watched) this talk very helpful, which help me answer some preliminary question and think deeper ones, just to answer my own previous question (missed the link just now) https://www.youtube.com/watch?v=R4EmE1QEvNU | 21:02 |
| openstackgerrit | Merged openstack-infra/nodepool: Add ZK API methods for node requests https://review.openstack.org/416671 | 21:43 |
| jeblair | i'm approving the rebased changes in the stack (carrying over reviews for ones that had rebase conflicts) | 21:44 |
| jeblair | pabelanger: you reviewed all the other changes in the stack, did you want to see https://review.openstack.org/417994 ? | 21:44 |
| pabelanger | jeblair: I likely skipped because my threads programming isn't the best. | 21:46 |
| pabelanger | but I can look now | 21:46 |
| openstackgerrit | Merged openstack-infra/nodepool: Disable tests for nodepoold changes https://review.openstack.org/417108 | 21:46 |
| openstackgerrit | Merged openstack-infra/nodepool: Remove Gearman and ZMQ https://review.openstack.org/417128 | 21:46 |
| openstackgerrit | Merged openstack-infra/nodepool: Minor code cleanup https://review.openstack.org/417199 | 21:47 |
| openstackgerrit | Merged openstack-infra/nodepool: Set valid states per data model class https://review.openstack.org/417482 | 21:48 |
| * SpamapS staring at merge branch code for the last 20 minute | 22:04 | |
| Shrews | jeblair: awesome. now i can put some thinking in how to limit the number of requests handled concurrently. | 22:09 |
| Shrews | thinking i'll have to track the threads spawned and compare alive count with the max | 22:10 |
| pabelanger | Shrews: 1 thread per vm that is launched right? | 22:11 |
| pabelanger | err, per provider | 22:12 |
| Shrews | pabelanger: well, 1 thread per provider, but each provider thread can spawn N threads to handle N requests | 22:12 |
| pabelanger | right | 22:12 |
| Shrews | we *might* want to put a limit on N | 22:13 |
| Shrews | i also suspect we might be hitting ZK pretty hard to traverse the request queue. that worries me a bit | 22:15 |
| jeblair | Shrews: we could get clever then and set a child watch on it and only traverse it when it updates? | 22:17 |
| jeblair | (could probably do that later to avoid premature optimization) | 22:17 |
| jeblair | (oh, but perhaps sometimes our own situation will change and we may want to revisit earlier requests...) | 22:18 |
| jeblair | forget i said anything until we have a better handle on the algorithm :) | 22:18 |
| * Shrews forgets all the things | 22:19 | |
| jeblair | Shrews: i'm seeing phrases like "15k requests per second" in relation to single zookeeper servers... | 22:19 |
| Shrews | hopefully that's not followed by "is waaaaay beyond the maximum" | 22:20 |
| Shrews | i may be worrying about nothing. i like to get the worrying out of the way early in the software development lifecycle | 22:20 |
| jeblair | Shrews: it was preceded by "is still able to process" :) | 22:21 |
| jeblair | look, a graph! https://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview | 22:21 |
| jeblair | and that includes almost half as write operations | 22:21 |
| jeblair | Shrews: what? me worry? | 22:22 |
| Shrews | well, i'm inclined to believe the Internet. sweeeet :) | 22:22 |
| Shrews | we'll have to remember to make our own shiny ZK graphs when this is done | 22:22 |
| * Shrews smells a talk topic | 22:22 | |
| jeblair | oh yeah, we should start shipping data over to graphite so we can grafana it | 22:22 |
| pabelanger | was looking at fedmsg again today | 22:26 |
| pabelanger | really need to get up a server to play with | 22:26 |
| * SpamapS trying to understand what actually lands the commits in the local git repos when using fakegerritconnection | 22:33 | |
| jeblair | SpamapS: FakeChange.setMerged | 22:35 |
| SpamapS | jeblair: ty.. just found that, then alt-tabbed back here to say that. ;) | 22:46 |
| SpamapS | I'm trying to re-enable test_build_configuration_branch and the commits seem to be overriding previous commits (so, A, B, and C, all land individually, and in the end, only C is in the branch) | 22:48 |
| *** hashar has quit IRC | 22:49 | |
| jeblair | SpamapS: i think that test is using the merger repo rather than the upstream repo to perform that validation -- it's checking that the repo state it prepared for the job including all 3. | 22:54 |
| jeblair | SpamapS: the fake gerrit isn't sophisticated enough to merge changes as needed (it only fast-forwards to whatever you tell it is merged) | 22:54 |
| jeblair | s/including/includes/ | 22:56 |
| SpamapS | jeblair: the fake gerrit seems to only be able to land one commit in the upstream repo. | 22:56 |
| jeblair | SpamapS: right, it just sets the pointer to that commit | 22:57 |
| SpamapS | it does in fact overrite whatever the previous head was, with the current change | 22:57 |
| SpamapS | overwrite even :-P | 22:57 |
| SpamapS | jeblair: ok, so that's just not behavior we can test, and the old test was able to rely on the merger keeping the ref.. | 22:58 |
| jeblair | SpamapS: right. we could make fakegerrit do that without too much trouble, but it may be easier to do the v3 equivalent of that, which is to check what's in the jobdir for the change. | 22:59 |
| SpamapS | jeblair: indeed, I'm looking at that right now | 23:00 |
| jeblair | SpamapS: if we grab a build for the last change, we should be able to do build.hasChanges(A,B,C) i think | 23:00 |
| jeblair | or something like that | 23:00 |
| SpamapS | oh that would be slick, but would that test the same thing? | 23:00 |
| jeblair | yeah, i think that would be as close to a straight port of the test as we can manage | 23:00 |
| jeblair | (i think the test would probably be improved by make fakegerrit more sophisticated and testing that, but that's probably best deferred) | 23:01 |
| jeblair | i feel like i butchered that sentence | 23:02 |
| jeblair | (i think the test would probably be improved by making fakegerrit more sophisticated and testing that changes landed upstream as expected, but that's probably best deferred) | 23:02 |
| jeblair | redo ^ | 23:02 |
| *** saneax-_-|AFK is now known as saneax | 23:20 | |
| Shrews | jeblair: mordred: pabelanger: can one of you +A https://review.openstack.org/418585? It has two +2's | 23:28 |
| jeblair | done | 23:30 |
| jeblair | Shrews: and parents | 23:31 |
| jeblair | Shrews: can you take a look at https://review.openstack.org/412160 ? | 23:31 |
| jeblair | (not urgent) | 23:31 |
| openstackgerrit | Merged openstack-infra/nodepool: Create per-provider ZK work threads https://review.openstack.org/417994 | 23:33 |
| openstackgerrit | Merged openstack-infra/nodepool: Remove --no-launches nodepoold option https://review.openstack.org/418001 | 23:34 |
| openstackgerrit | Merged openstack-infra/nodepool: Remove --no-builder nodepoold option https://review.openstack.org/418013 | 23:35 |
| Shrews | jeblair: thanks, and yes. first thing in the morning (or after making dinner tonight) | 23:35 |
| openstackgerrit | Merged openstack-infra/nodepool: Handle provider removal from config https://review.openstack.org/418112 | 23:35 |
| openstackgerrit | Merged openstack-infra/nodepool: Add framework for handling node requests https://review.openstack.org/418585 | 23:35 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!