Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-04-20
11:45:07 openstackgerrit Surya Seetharaman proposed openstack/nova master: Delete orphaned RS and IM records of archived instances https://review.openstack.org/560042
11:45:43 openstackgerrit Nguyen Hai proposed openstack/nova-specs master: Follow the new PTI for document build https://review.openstack.org/551802
11:46:15 cdent no, not really. I'm not much of a customizer. And if it wasn't there already in a sort of semi-automatic fashion, I'd not turn on the syntax coloring. The main things I care about are auto-indent and paren matching. And since nova uses such a bad indent style, I'm not getting much help from my editor on that front (I could fix my editor to behave better, but haven't got around to it)
11:49:10 fried_rice cdent: I love vim just fine, but IDE ability to jump from a symbol to its definition, find all usages of a method, jump directly to method overrides, search for symbols by (wildcarded) name, see class/method/attribute overview trees, etc. -- all pretty serious productivity boosters for me.
11:49:28 cdent tags
11:50:01 fried_rice There was a time I used to be able to do interactive debugging, but it didn't (used to) work in the nova project, and I haven't tried it recently or tried to fix it.
11:50:24 fried_rice Do you tag the whole project?
11:50:47 cdent fried_rice: do you ever wonder or worry if IDE superpowers can result in code that is harder for other people to read/maintain?
11:51:08 cdent when I use them, yes, but I frequently don't
11:51:14 cdent because I don't need/want to
11:51:51 fried_rice cdent: You mean because I'd be less inclined to write new stuff in sensible locations because I know I can always navigate by hotkey?
11:52:37 cdent not just that, not just about navigation, and not really anything in particular: more just "do you think about that?"
11:52:41 fried_rice Given formatting restrictions (which pycharm also validates on the fly) I'm not sure what other pitfalls there might be.
11:52:52 cdent me neither, just noodling
11:53:11 fried_rice And given our review process, I would expect people to gripe if stuff was in the wrong place.
11:53:38 cdent Yeah, I'm not really thinking about issue of where code is
11:53:55 cdent More about the actual code itself: how it is written, how it is structured internally to itself
11:54:23 fried_rice Mm. No, I don't feel like I would do any of that differently in a primitive editor.
11:54:56 fried_rice If code generation were a thing, I could see that being an issue (though I would also expect that to get filtered by reviewers)
11:55:19 fried_rice So like, if this was Java, and I was using "generate a bean" or whatever.
11:56:07 fried_rice But - whether because I just don't know about / use the code generation features, or they don't exist, or they wouldn't apply to the kind of dev we do in openstack, or whatever - I don't use any code gen tools in my IDE.
11:56:59 fried_rice btw, I tried quick-building that doc and apparently my venv wasn't set up, so I'm going through tox.... and the venv is *still* installing deps.
11:57:22 fried_rice Which says to me that we're installing the wrong set of deps for building docs.
11:58:04 fried_rice I think stephenfin has some patches that clean some of that up. But I think they just consolidate some of the venvs. Not sure whether docs was folded into one of the others.
11:58:37 cdent the gate now builds docs outside of the 'docs' venv that can be used locally
11:59:00 cdent but yes, the deps are probably wrong. probalby building nova deps, not docs deps
12:04:38 openstackgerrit sahid proposed openstack/nova master: libvirt: don't set interface backend for vhostuser https://review.openstack.org/562708
12:04:39 openstackgerrit sahid proposed openstack/nova master: libvirt: fix wrong driver name for vhostuser interface https://review.openstack.org/561562
12:04:40 openstackgerrit sahid proposed openstack/nova master: libvirt: add support for virtio-net rx/tx queue sizes https://review.openstack.org/484997
12:06:27 sahid giblet: i added a test i hope it's what you were looking for
12:11:15 frickler melwitt: gah, if you had created a bug for https://review.openstack.org/561370 it might have saved me searching half a day to come to the same result. but then maybe not, so no hard feelings ;)
12:11:55 openstackgerrit Yikun Jiang (Kero) proposed openstack/nova master: WIP: complex policy https://review.openstack.org/553776
12:12:19 openstackgerrit sahid proposed openstack/nova master: libvirt: fix wrong driver name for vhostuser interface https://review.openstack.org/561562
12:12:20 openstackgerrit sahid proposed openstack/nova master: libvirt: add support for virtio-net rx/tx queue sizes https://review.openstack.org/484997
12:12:51 neha_alhat stephenfin: Hi
12:16:08 bauzas folks, I'm half-around
12:16:22 bauzas my wife is in ER, so I'll help her
12:17:07 fried_rice bauzas: Hope everything is okay
12:17:14 neha_alhat stephenfin: Actually I want to know why these two conf parameters regrading live migration are made mutable in[1]. [1]:https://review.openstack.org/#/c/319203/
12:19:47 bauzas fried_rice: hope so
12:22:08 openstackgerrit Yikun Jiang (Kero) proposed openstack/nova master: WIP: complex policy https://review.openstack.org/553776
12:27:42 stephenfin fried_rice: Yeah, I need to pick those up again and rebase them
12:28:03 stephenfin Especially now, given the whole "move everything to python3" initiative
12:28:04 stephenfin Also
12:28:31 finucannot neha_alhat: What do you mean?
12:28:54 openstackgerrit Jens Harbott (frickler) proposed openstack/nova master: DNM: Test devstack + ceph plugin patches https://review.openstack.org/563097
12:28:56 fried_rice finucannot: Do you happen to know offhand why a doc containing a `link`_ with a missing _`target` would have passed sphinx build?
12:29:13 fried_rice Is that warning not fatal in the nova repo?
12:29:24 fried_rice I thought it was, but maybe I'm thinking of the nova-specs repo
12:29:45 finucannot fried_rice: I bet it fails locally https://review.openstack.org/#/c/559348/
12:30:32 fried_rice oh, I remember seeing that patch.
12:30:51 neha_alhat finucannot: why only these parameters are marked mutable=True? why not others? Is there any criteria to make conf parameters as mutable?
12:31:00 finucannot fried_rice: Yeah, ima go bug infra guys and see if we can get some movement on it
12:31:05 fried_rice finucannot: But so the answer is that the gate won't fail on those for now?
12:31:23 finucannot fried_rice: Correct, and there's nothing we can do to override that
12:31:32 finucannot Save a custom extension or something
12:32:41 finucannot neha_alhat: Things are generally mutable if you it's something that you feel you'd want to change at runtime (and actually can without breaking stuff)
12:32:49 finucannot I guess this was one of those things
12:33:12 finucannot neha_alhat: tbh, I don't have the full context and the commit message is poor so I can't really say what the authors intention was
12:33:31 finucannot This is why one should always write mdbooth-level commit messages :)
12:35:10 fried_rice finucannot: I'm +2ing cdent's https://review.openstack.org/#/c/563086/ - after having checked the rendered doc because apparently I can't rely on sphinx for this one...
12:35:46 finucannot fried_rice: I've +Wd it
12:35:52 fried_rice ack
12:36:21 fried_rice That'll teach cdent to use the word "trivial".
12:37:06 neha_alhat finucannot:Do you have any idea if we mark parameter as muatble=True, then it takes imediately the modified value or it requires sighup signal to pass to load that modified value?
12:38:16 finucannot neha_alhat: Docs say the latter https://docs.openstack.org/oslo.config/latest/reference/mutable.html
12:42:06 openstackgerrit Eric Fried proposed openstack/nova master: Remove deprecated [placement] opts https://review.openstack.org/557086
12:42:34 fried_rice finucannot: re+W after rebase, please? ^
12:43:00 finucannot fried_rice: sure
12:43:02 fried_rice thx
12:44:05 giblet sahid: thanks for the unit test
12:46:20 neha_alhat finucannot: In the above docs it is mentioned that: after making conf parameter mutable, they can be reloaded (usually via SIGHUP) at runtime, without a service restart.
12:48:53 neha_alhat finucannot: Actually from the patch[1], what I understood is: If the conf parameter value is cached outside loop and if it is getting used in loop for long running processes like live migration, if operator changes parameter value then it should imediately get reflected
12:49:19 neha_alhat finucannot: [1]: https://review.openstack.org/#/c/319203/
12:50:39 fried_rice finucannot: I'm having trouble getting my docs venv to build locally (I killed it after like 20 minutes - should it take that long??) so while I try to wrangle that, do you happen to know off hand what's going on here? http://logs.openstack.org/57/517757/13/check/build-placement-api-ref/b43304e/job-output.txt.gz
12:51:08 fried_rice clearly a bug in the tooling, if only because it's not showing me the actual error...
12:51:36 finucannot neha_alhat: I'm not the person to ask, to be honest. The best thing to do would be build a minimal test case and see what happens. If it works as you expect, a patch to the oslo.config docs would be really helpful
12:53:18 neha_alhat finucannot: thanks for input
12:53:33 finucannot fried_rice: That looks like you're assigning a string to something os-api-ref expects to be a dict
12:53:49 fried_rice in a doc?
12:53:57 finucannot In a YAML file, I'd suspect
12:54:00 fried_rice o
12:54:06 finucannot So...this? https://review.openstack.org/#/c/517757/13/placement-api-ref/source/parameters.yaml
12:54:06 fried_rice ...
12:54:36 finucannot Lemme pull it down and try
12:57:01 fried_rice finucannot: Yeah, I can't see anything bizarre in there.
12:57:33 fried_rice Should I open a bug about that crazy error message? (Although you seem not to have had trouble deciphering it)
12:58:06 finucannot fried_rice: Yeah, some additional error messages wouldn't be any harm, I imagine
13:03:41 finucannot fried_rice: Yup, you'd a typo. Added comment
13:03:47 finucannot I'll go fix os-api-ref now
13:03:50 fried_rice finucannot: Thanks!
13:04:41 finucannot fried_rice: Did you end up opening a bug? If not, I'll just send this patch as-is
13:05:05 fried_rice finucannot: I'm in the process. Two shakes.
13:05:53 fried_rice finucannot: https://github.com/sphinx-doc/sphinx/issues/4869
13:06:13 finucannot fried_rice: Ah, that's an os-api-ref bug. Not a Sphinx bug
13:06:36 fried_rice finucannot: I was just following the instructions in the error output.
13:06:38 finucannot fried_rice: You can tell based on this bit --> File "/home/zuul/.venv/local/lib/python2.7/site-packages/os_api_ref/__init__.py", line 261, in _check_yaml_sorting
13:07:01 finucannot Yeah, that error message is dumb. _That's_ probably a good Sphinx bug
13:08:59 fried_rice finucannot: Should I close that one, then?
13:09:07 finucannot Probably best, yeah

Earlier   Later