Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-05-05
20:42:35 cdent please please sdague email _discussions_ ;)
20:42:42 cdent why should you resist mriedem
20:42:42 sdague many people find the random flaming basically too emotionally sapping to want to participate in discussions at all, so what happens is all the thoughtful middle gives up
20:42:50 mriedem i'm glad i wasn't the only one that interpreted "we need you to do LTS because we have a hard time keeping up with our forks"
20:43:14 dansmith at least oracle funds so many people working on stuff in the community
20:43:44 mriedem to date, there is one person from oracle helping with multiattach
20:44:05 mriedem i will give credit where credit is due
20:44:15 cdent sdague: I guess I'm weird: I'm so conscious of the under- or un-expressed emotion that that is the thing that I find so sapping.
20:44:28 cdent incredibly so
20:45:11 sdague cdent: sure, that might be you. Lots of people basically end up ignoring the whole of the mailing list for some number of weeks after this because they see the mailing list as a place where loud people yell about things.
20:45:43 cdent do you consider the current conversation to be yelling already or not quite there yet?
20:45:56 sdague I think it is just about there
20:45:57 mriedem stvnoyes: btw, let me thank you personally for helping out with the multiattach stuff
20:46:27 sdague cdent: which is why I "save draft" on my last email
20:46:29 mriedem stvnoyes: also gave you a tip in https://review.openstack.org/#/c/456877/6/nova/compute/manager.py
20:46:47 cdent wow. I guess I spent too much time on usenet in the 90s
20:46:51 sdague cdent: yep
20:47:35 cdent to me this feel like polite but earnest dinner conversation
20:47:57 sdague I think if you polled the audience you would not get that reaction
20:50:14 melwitt mriedem: cool, I figured. since we're going to be at the summit anyway, we could chat briefly about it then
20:50:32 mriedem melwitt: perhaps on thursday night
20:50:45 mriedem which is the only night i'll get to enjoy the company of my fellow nova compatriots
20:52:00 melwitt mriedem: k. bummer we only get to hang with you one evening
20:52:12 mriedem and sunday night
20:52:13 mriedem and friday morning
20:52:22 mriedem and the 2 minutes in between sessions
20:52:29 efried mriedem Would you mind pushing https://review.openstack.org/#/c/391288/ ? I'm getting merge conflicts up the chain, and would like to limit how much rebasing is needed.
20:53:23 mriedem efried: don't you know i'm all riled up?
20:53:28 mriedem :)
20:53:37 efried mriedem Me too. The day I've had.
20:53:53 mriedem just blame it on thorst_afk
20:53:57 efried Always.
20:54:21 thorst_afk he's always afk
20:54:38 melwitt hah
20:56:16 efried Honestly, I'm not sure how https://review.openstack.org/#/c/443189/ got a merge conflict but its predecessors didn't.
20:56:32 efried It didn't touch any non-PowerVM files.
20:57:32 mriedem https://review.openstack.org/#/c/456877/7/nova/tests/unit/compute/test_compute.py@10260
20:57:32 mriedem melwitt: i'm suprised this doesn't result in a keyerror
20:57:36 mriedem since it's not using a kwarg
20:57:58 mriedem oh i know why
20:58:02 mriedem it's using attachment_id as the message
20:58:55 melwitt mriedem: yeah, that might match because of python named argument matching?
20:59:08 mriedem no i think it's because
20:59:08 mriedem def __init__(self, message=None, **kwargs):
20:59:12 mriedem if not message:
20:59:23 mriedem if you overwrite the message, we don't try the msg % kwargs
20:59:26 mriedem and get the KeyError
20:59:37 melwitt oh, I see.
21:00:00 openstackgerrit Matt Riedemann proposed openstack/nova master: Add Cinder v3 detach to shutdown_instance https://review.openstack.org/456877
21:02:03 melwitt I was thinking of the named argument thing but that works in the opposite way. you can specify positional args as kwargs by name (and in wrong order) and it will still work
21:02:56 melwitt s/positional/normal/
21:03:47 efried aHA!
21:03:51 efried It's melwitt's fault!
21:04:14 efried https://review.openstack.org/#/c/462764/3/nova/tests/unit/virt/powervm/test_driver.py
21:04:33 melwitt lol, you meant literally
21:04:34 mriedem heh, well, it's really your fault
21:05:14 mriedem stvnoyes: btw, reverify is ignored
21:05:15 mriedem use recheck
21:06:07 efried So, hm, is there any way for me to fix my merge conflict without rebasing the predecessors?
21:18:34 thingee hey anyone oppose me proposing the deprecation of the trusted filter?
21:18:54 thingee it has been experimental since it has existed in may 28 2012
21:20:06 cdent new things crawl out of the woodwork every day
21:20:52 thingee aight, I'll send an email to ops list to see if anyone cares.
21:21:00 cdent good plan
21:22:34 thingee http://lists.openstack.org/pipermail/openstack-operators/2017-May/013371.html boom
21:39:31 stvnoyes mriedem: ok thanks. I had read somewhere that reverify was used to re-kick a gate test, but it must have been stale info
22:20:47 openstackgerrit Mike Perez proposed openstack/nova master: Deprecate scheduler trusted filter https://review.openstack.org/463065
22:24:08 openstackgerrit Mike Perez proposed openstack/nova master: Deprecate scheduler trusted filter https://review.openstack.org/463065
22:42:55 mordred sdague: I feel like maybe I'm skipping too much of that thread
22:51:01 mordred sdague: oh - nope. reading more of it does not make my life any better
23:11:51 mriedem thingee: that reminds me https://review.openstack.org/#/c/457812/
23:12:09 mriedem stvnoyes: in the long ago reverify would do the same thing as recheck, but i think they stopped handling that a few months ago
23:12:39 thingee mriedem: deprecate all the filters
23:13:19 mriedem thingee: btw, what prompted this?
23:18:53 thingee The board/tc meeting discussed a goal of simplification. Reduction of options and what not
23:18:56 thingee mriedem: ^
23:19:28 thingee I was pointed to this when I was gathering ideas
23:20:12 thingee mriedem: let me know if you have suggestion of things worth looking into
23:20:20 mriedem thingee: holy crap we probably have a ton
23:20:29 mriedem thingee: i know we deprecated a lot of things in ocata,
23:20:40 mriedem i think you could just be looking for things that are deprecated for removal
23:21:02 mriedem assuming someone doesn't already have a patch up to remove those things and we just haven't seen it yet
23:21:07 thingee mriedem: yeah I removed something out of nova recently that was marked deprecated. not sure if I saw anything else, but I'll double check
23:23:50 mriedem thingee: grep for deprecated_since='15.0.0', and you'll see a few things
23:24:06 mriedem although,
23:24:21 mriedem i'd avoid anything under nova/conf/network as those are probably for nova-network which isn't gone yet
23:24:39 mriedem really anything that's deprecated related to nova-network is just signaling right now
23:25:03 mtreinish mriedem: an even longer time ago reverify would just run the gate queue again and not rerun check
23:25:09 mriedem anyway, got to run
23:25:11 mriedem mtreinish: i know
23:25:12 mriedem sheesh
23:25:40 mtreinish thingee: also you can look at logstash we always emit a ton of deprecation warnings during gate jobs: http://logstash.openstack.org/#/dashboard/file/logstash.json?query=message:%5C%22deprecated%5C%22%20AND%20loglevel:%5C%22WARNING%5C%22%20AND%20build_branch:%5C%22master%5C%22
23:25:59 mriedem mtreinish: i didn't want to expose my nerd-like internal knowledge about the slight differences between recheck and reverify and the check vs gate queues :)
23:26:04 mtreinish thingee: I made a push to try and clean that up a while ago, but it never really fixed the problem: http://lists.openstack.org/pipermail/openstack-dev/2015-December/081984.html
23:26:23 mtreinish mriedem: heh, fair enough
23:26:36 mriedem mtreinish: btw, you still owe me that cloudpipe cli/api deprecation patch for novaclient
23:26:46 mtreinish oh, I do don't I
23:26:54 mtreinish I'll look at that in a couple min
23:27:05 mriedem i'm gonna go gamble
23:27:46 mtreinish enjoy

Earlier   Later