| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-09-22 | |||
| 00:55:06 | mriedem | ok left some comments inline, you can add docs to it in the end if you want | |
| 00:55:57 | openstackgerrit | Dan Smith proposed openstack/nova master: Fix minor input items from previous patches https://review.openstack.org/506416 | |
| 00:55:59 | dansmith | mriedem: ^ | |
| 00:57:04 | dansmith | mriedem: does that help at all? or is it still too obscure because I'm not comparing against a static obvious list? | |
| 00:58:42 | mriedem | you only updated the commit message | |
| 00:59:07 | dansmith | gah | |
| 00:59:27 | openstackgerrit | Dan Smith proposed openstack/nova master: Fix minor input items from previous patches https://review.openstack.org/506416 | |
| 00:59:31 | dansmith | sorry | |
| 01:01:08 | mriedem | ok, left a comment about the key_name thing in the last change | |
| 01:02:39 | openstackgerrit | Dan Smith proposed openstack/nova master: Fix minor input items from previous patches https://review.openstack.org/506416 | |
| 01:02:54 | dansmith | got it | |
| 01:03:30 | mriedem | i think i'm getting the global marker patch now | |
| 01:03:37 | mriedem | understanding the patch before it is key | |
| 01:10:30 | mriedem | so for the ip filtering we do in python in the compute api, does the end of this result in us returning a sorted/filtered list to the code that then processes the ip filter regex? | |
| 01:15:03 | dansmith | mriedem: I'm not sure I understand.. that filtering code is in the api, right? we shouldn't be changing what it sees | |
| 01:15:14 | mriedem | yeah i just haven't gotten there yet | |
| 01:15:22 | dansmith | okay | |
| 01:15:27 | mriedem | this is all basically a shim in between the db api and that code that processes the result and filters by ip regex | |
| 01:15:28 | mriedem | ok | |
| 01:15:36 | dansmith | right | |
| 01:15:49 | dansmith | a multiplexing layer that doesn't eff up the sort order like what we had before | |
| 01:20:04 | mriedem | ok just some nits in the mega marker paging patch, which can go at the end, otherwise +2 | |
| 01:20:07 | mriedem | my work here is done | |
| 01:20:20 | dansmith | cool thanks I'll take a look in a sec | |
| 01:33:19 | openstackgerrit | Dan Smith proposed openstack/nova master: Fix minor input items from previous patches https://review.openstack.org/506416 | |
| 01:49:22 | dansmith | melwitt: they work with desc too, will push a test variant at the end | |
| 01:49:41 | mriedem | hey, don't forget to read the ML before bed | |
| 01:49:48 | mriedem | there is a surprise | |
| 01:50:20 | openstackgerrit | Dan Smith proposed openstack/nova master: Fix minor input items from previous patches https://review.openstack.org/506416 | |
| 01:50:28 | dansmith | mriedem: the mogan stuff? | |
| 01:50:32 | dansmith | melwitt: ^ | |
| 01:50:42 | mriedem | no | |
| 01:50:44 | mriedem | better | |
| 01:50:50 | mriedem | an old set of friends | |
| 01:50:54 | mriedem | coming together at last | |
| 01:51:09 | melwitt | dansmith: cool, thanks | |
| 01:52:31 | melwitt | I was thinking between the tests earlier in the series to now, everything stitches together, but wasn't totally sure | |
| 01:52:59 | dansmith | melwitt: yeah, I think it's all the same enough that it wouldn't matter, but good to have them in there | |
| 01:54:37 | melwitt | lol mriedem | |
| 01:55:22 | dansmith | oh, the shelve and file injection thing? | |
| 01:55:28 | mriedem | yeah | |
| 01:55:42 | mriedem | oh come on | |
| 01:55:53 | melwitt | I chuckled | |
| 01:55:56 | mriedem | shelve is one of my most favorite things to shit on in the ML on a semi-monthly basis | |
| 01:56:05 | mriedem | melwitt: you said you "lol"ed | |
| 01:56:14 | melwitt | I chuckled out loud | |
| 01:56:28 | mriedem | ok | |
| 01:56:29 | melwitt | a chuckle is a laugh right? | |
| 01:56:32 | dansmith | issubclass('chuckle', 'laugh') == True | |
| 01:56:36 | mriedem | ok | |
| 01:56:41 | melwitt | haha yeeeah | |
| 01:56:41 | mriedem | didn't know the rules | |
| 01:56:48 | dansmith | rules ^ | |
| 01:56:58 | mriedem | you know, | |
| 01:57:09 | mriedem | if we allowed you to just specify personality files on unshelve... | |
| 01:57:13 | mriedem | basically, | |
| 01:57:26 | dansmith | you know, if we took shelve and file injection out back and set them on fire... | |
| 01:57:29 | mriedem | if we just allowed you to specify everything you can specify on initial create, but on rebuild and unshelve, we'd be good | |
| 01:57:42 | mriedem | well we're doing that with file injection this release, i just need to spec that up | |
| 01:57:42 | melwitt | well yeah, that's what everyone wants | |
| 01:58:01 | mriedem | maybe shelve/unshelve should be a topic at the forum | |
| 01:58:06 | mriedem | "can we please just remove this? please?" | |
| 01:58:25 | dansmith | getting input on who actually uses it and for what seems fairly legit to me, fwiw | |
| 01:58:40 | dansmith | like if nobody cares if it keeps working, I care a lot less than I already do | |
| 01:58:54 | dansmith | and I care very little | |
| 01:59:05 | mriedem | yeah we still kick that idea around on figuring out who is using the APIs | |
| 01:59:08 | mriedem | maybe that's a brainstormer | |
| 01:59:48 | melwitt | based on the amount of bugs I see with it, I will guess a fair number of people use it | |
| 01:59:56 | melwitt | *bug reports | |
| 02:00:28 | dansmith | I want to know how many people use it for real, vs. use tempest to verify their cloud and continually find it broken and think it's worth reporting | |
| 02:00:45 | dansmith | but, also getting 'why do you want to use this' answered is useful I think | |
| 02:00:53 | dansmith | because if they're using it instead of snapshot for dumb reasons... | |
| 02:00:56 | melwitt | yeah. just musing | |
| 02:01:34 | melwitt | talkin it up | |
| 02:01:38 | melwitt | talkin bout topics | |
| 02:03:15 | mriedem | the tempest tests are super simple too | |
| 02:04:25 | openstackgerrit | wanghongtaozz proposed openstack/nova master: Fix typos https://review.openstack.org/506437 | |
| 02:04:47 | dansmith | oooh, typos | |
| 02:05:03 | dansmith | fix _all_ the typos | |
| 02:06:19 | mriedem | -1'ed with fury | |
| 02:07:22 | mriedem | https://review.openstack.org/#/q/owner:wanghongtaozz+status:open | |
| 02:07:23 | mriedem | a nice pattern | |
| 02:08:07 | melwitt | ... yeah | |
| 02:17:52 | openstackgerrit | Merged openstack/nova master: Add db.instance_get_by_sort_filters() https://review.openstack.org/504985 | |
| 02:27:39 | openstackgerrit | Matt Riedemann proposed openstack/nova master: WIP: Remove allocations when unshelve fails on host https://review.openstack.org/506458 | |
| 03:17:48 | alex_xu | dansmith: just found the default dir is 'desc', not the 'asc', the confusing thing is we pass a different different_dir to process_sort_params...https://review.openstack.org/#/c/504983/8/nova/compute/instance_list.py@87 | |
| 03:19:13 | dansmith | alex_xu: I don't think it matters anyway, right? | |
| 03:20:30 | dansmith | some default sort order helps for consistency, but if the user didn't ask for something specific I don't think we have any requirement to give them something specific | |
| 03:22:57 | alex_xu | dansmith: yes, the only thing is the default order changed, the API behaviour should be stable? | |
| 03:24:04 | dansmith | alex_xu: if you think it's important we can change it to desc by default, but I don't think it matters | |
| 03:24:18 | dansmith | if you don't ask for a sort order, I don't think you can complain about which way we sort it :) | |
| 03:24:34 | dansmith | sorting desc by default also seems weird to me, but.. | |
| 03:26:31 | alex_xu | dansmith: emm...only can complain is the api-ref talk about default order https://developer.openstack.org/api-ref/compute/#id4 :) | |
| 03:27:16 | dansmith | alex_xu: where? I don't see "sort" or "order" on that page | |
| 03:27:42 | dansmith | the default sort includes db id, so I hope we're not saying that in our api ref :) | |
| 03:28:12 | alex_xu | dansmith: the description of sort_key, "Sorts by a server attribute. Default attribute is created.", but yea, there is mistake, it is created_at, not created | |
| 03:28:27 | alex_xu | dansmith: yea, good luck, we didn't mention that :) | |
| 03:29:02 | dansmith | alex_xu: I don't see where you see that in that doc | |
| 03:29:30 | alex_xu | dansmith: the request parameters of https://developer.openstack.org/api-ref/compute/#list-servers | |
| 03:30:00 | dansmith | okay I see it | |