| *** saneax is now known as saneax-_-|AFK | 00:01 | |
| *** Cibo_ has joined #zuul | 05:01 | |
| *** bhavik1 has joined #zuul | 05:31 | |
| *** bhavik1 has quit IRC | 05:38 | |
| *** auggy has quit IRC | 07:28 | |
| *** TheJulia has quit IRC | 07:29 | |
| *** bstinson has quit IRC | 07:30 | |
| *** TheJulia has joined #zuul | 07:34 | |
| *** bstinson has joined #zuul | 07:34 | |
| *** auggy has joined #zuul | 07:34 | |
| *** Cibo_ has quit IRC | 08:03 | |
| *** yolanda has quit IRC | 08:18 | |
| *** yolanda has joined #zuul | 08:19 | |
| *** hashar has joined #zuul | 09:44 | |
| *** openstackgerrit has quit IRC | 10:18 | |
| *** yolanda has quit IRC | 10:23 | |
| *** yolanda has joined #zuul | 10:24 | |
| *** yolanda has quit IRC | 10:26 | |
| *** yolanda has joined #zuul | 10:27 | |
| *** yolanda has quit IRC | 10:39 | |
| *** yolanda has joined #zuul | 10:41 | |
| *** yolanda has quit IRC | 10:46 | |
| *** yolanda has joined #zuul | 10:49 | |
| *** yolanda has quit IRC | 10:49 | |
| *** yolanda has joined #zuul | 10:49 | |
| *** hashar has quit IRC | 13:35 | |
| *** bhavik1 has joined #zuul | 13:59 | |
| *** bhavik1 has quit IRC | 14:24 | |
| *** yolanda has quit IRC | 14:50 | |
| *** yolanda has joined #zuul | 14:50 | |
| *** yolanda has quit IRC | 14:54 | |
| *** yolanda has joined #zuul | 14:54 | |
| *** yolanda has quit IRC | 14:55 | |
| *** yolanda has joined #zuul | 14:56 | |
| *** yolanda has quit IRC | 14:57 | |
| *** yolanda_ has joined #zuul | 14:59 | |
| *** yolanda_ has quit IRC | 14:59 | |
| *** yolanda_ has joined #zuul | 14:59 | |
| *** yolanda__ has joined #zuul | 15:00 | |
| *** yolanda_ has quit IRC | 15:01 | |
| *** yolanda__ is now known as yolanda | 15:01 | |
| Shrews | jeblair: I'm about to put up a review that might help with that "no limit on threads" problem that you were worried about. Feedback welcome. | 16:18 |
|---|---|---|
| *** openstackgerrit has joined #zuul | 16:18 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add support for max_concurrency for providers https://review.openstack.org/420057 | 16:18 |
| Shrews | ^^^ | 16:18 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add support for max_concurrency for providers https://review.openstack.org/420057 | 16:32 |
| Shrews | ugh, bugs | 16:32 |
| *** yolanda has quit IRC | 16:35 | |
| Shrews | ugh, more bugs | 16:39 |
| *** yolanda has joined #zuul | 16:42 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add support for max_concurrency for providers https://review.openstack.org/420057 | 16:45 |
| pabelanger | I know everybody loves bikeshedding, but would max-workers be a better name of the setting? max-concurrency seems a little vague | 16:48 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add support for max_concurrency for providers https://review.openstack.org/420057 | 16:50 |
| Shrews | pabelanger: i don't care, personally. whatever folks prefer is fine with me | 16:51 |
| *** yolanda has quit IRC | 16:51 | |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add support for max_concurrency for providers https://review.openstack.org/420057 | 16:52 |
| Shrews | ok, that was the last patch set :) | 16:52 |
| Shrews | kept finding things to clean up | 16:53 |
| *** yolanda has joined #zuul | 16:59 | |
| *** yolanda has quit IRC | 17:04 | |
| *** yolanda has joined #zuul | 17:08 | |
| *** Shuo_ has joined #zuul | 17:13 | |
| Shuo_ | Is there a place on the dashboard seeing the average build/test time for jobs? | 17:14 |
| clarkb | Shuo_: zuul does predicted times but that is the closest you will get from the zuul status page I think. graphite does track runtimes though so the data can be retrieved from there | 17:15 |
| Shuo_ | clarkb: There are two numbers about the time in each status card on the dashboard (when I am looking at it now), is the 'predicted time' the one on top or the bottom? | 17:19 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume Task and TaskManager from shade https://review.openstack.org/414759 | 17:20 |
| clarkb | Shuo_: top is prediected time remaining, bottom is total time spent in zuul | 17:23 |
| Shuo_ | clarkb: Thanks. Could you give me a pointer (or a graphite config link), showing a concrete example? Such as 'the average test job time for a Nova commit" | 17:27 |
| clarkb | Shuo_: I don't think such a thing exists, what we do have are graphs like the job runtime graphs at the bottom of http://grafana.openstack.org/dashboard/db/nodepool those are for specific jobs mapped onto cloud providers | 17:28 |
| Shuo_ | clarkb: ok, good to know, thanks. What I wanted to know is 'how long a typical test job will be taking in my build farm'. If we can record such number, then we can do a lot of capacity forcast related things (e.g., combining that number with the # of job waiting in the queue, we can calculate things like 'by adding X more nodes, we can reduce the expected waiting time to Y') | 17:34 |
| clarkb | Shuo_: yup and that data is there, you just have to ask graphite for it | 17:35 |
| Shuo_ | clarkb: query the api directly? | 17:35 |
| clarkb | yes | 17:35 |
| pabelanger | clarkb: thanks for the additional summarize on v3 shim, I think I now understand the issue at hand. | 17:40 |
| openstackgerrit | David Shrewsbury proposed openstack-infra/nodepool: Add support for max_concurrency for providers https://review.openstack.org/420057 | 17:57 |
| *** bhavik1 has joined #zuul | 18:01 | |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume Task and TaskManager from shade https://review.openstack.org/420118 | 18:07 |
| *** bhavik1 has quit IRC | 18:07 | |
| *** bhavik1 has joined #zuul | 18:07 | |
| *** bhavik1 has quit IRC | 18:33 | |
| mordred | Shuo_: we don't have that specific thing graphed because we have jobs with WILDLY different times (our pep8 jobs, for instance, take maybe 2 minutes, our unittest jobs between 2 and 30 and our devstack integration jobs between 30 and 60 minutes) | 18:48 |
| Shuo_ | mordred: I was asking if the zuul pipeline has such capability to display (and retain for long term trend analysis) such data, and it sounds like it has. Specifically, if we take a deeper look at "zuul job queue" graph, we see sometimes job-in-wait number is quite high (we'd like to know if adding more capacity would help and what's the marginal ROI for added servers) | 18:54 |
| *** openstack has joined #zuul | 19:02 | |
| clarkb | indeed. Keep in mind though that its not purely a number of resources problem | 19:02 |
| clarkb | slow tests will be slow regardless of how many other instances you have available | 19:02 |
| clarkb | and that creates a lower bound on the runtimes | 19:03 |
| openstackgerrit | Monty Taylor proposed openstack-infra/nodepool: Consume Task and TaskManager from shade https://review.openstack.org/414759 | 19:05 |
| *** jamielennox is now known as jamielennox|away | 19:12 | |
| Shuo_ | mordred: agreed, it's not pure resource problem. | 19:12 |
| *** openstack has joined #zuul | 20:01 | |
| *** Shuo_ has quit IRC | 20:19 | |
| *** openstack has joined #zuul | 20:31 | |
| *** jamielennox|away is now known as jamielennox | 21:59 | |
| *** Shuo_ has joined #zuul | 22:05 | |
| *** Shuo_ has quit IRC | 22:12 | |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool: Move to fedora-25 for devstack testing https://review.openstack.org/420197 | 22:25 |
| openstackgerrit | Paul Belanger proposed openstack-infra/nodepool: Move to fedora-25 for devstack testing https://review.openstack.org/420197 | 22:29 |
| openstackgerrit | Jesse Keating proposed openstack-infra/zuul: Add trigger capability on github pr review https://review.openstack.org/419684 | 23:55 |
| *** harlowja has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!