Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2017-05-30
17:12:05 edleafe bauzas: ok, had a few minutes between meetings. Looking at the difference between my last patch and yours (https://review.openstack.org/#/c/465171/7..9), the only change I see is the cell0 context change
17:12:07 jaypipes rfolco: ^^ just fixed a typo and repushed.
17:12:20 edleafe bauzas: that doesn't appear to be related to this patch at all
17:12:33 edleafe bauzas: shouldn't it be a separate patch?
17:26:52 efried jaypipes Got a minute?
17:28:26 jaypipes efried: can you give me about 30 mins to review your get_inventory() patch? :)
17:29:01 openstack bug 1689692 in OpenStack Compute (nova) "MarkerNotFound when limit>num(instances) and marker starts in cell0" [High,In progress] https://launchpad.net/bugs/1689692 - Assigned to Matt Riedemann (mriedem)
17:29:01 openstackgerrit Merged openstack/nova master: Add recreate functional test for regression bug 1689692 https://review.openstack.org/468549
17:29:07 efried jaypipes Absolutely, thanks. I think I missed you on Friday, but here are the questions/uncertainties I had: http://eavesdrop.openstack.org/irclogs/%23openstack-nova/%23openstack-nova.2017-05-26.log.html#t2017-05-26T19:55:33
17:29:57 jaypipes efried: heh. "claimified" :) nice.
17:30:06 efried (TM)
17:30:45 jaypipes efried: :) so, the answer to your question #1 is yes, exactly. we handle allocations totally separately now.
17:31:10 jaypipes efried: as for the second question, can you give me examples of these resource classes that will not be consumed?
17:31:18 efried k. That's gonna make 'reserved' way more important (and harder to handle for us).
17:31:23 sdake is there any recent regression in the nova client or nova itself in master?
17:31:41 efried jaypipes I'm not sure they were being consumed in the first place. There's a link: https://github.com/openstack/nova/blob/master/nova/virt/powervm/host.py#L55-L64
17:31:59 efried I noticed similar resources from get_available_resource in one of the other drivers - ironic, I think.
17:32:06 jaypipes efried: oh, yeah, those aren't resources. :)
17:32:12 efried Which, as far as I can tell, are also going to be ignored under get_inventory
17:32:42 jaypipes well, the numa_topology may be a resource in the (hopefully near) future. but things like hypervisor_type and supported_instances, etc, are not resource classes.
17:32:59 jaypipes efried: yes, those are ignored by get_inventory()
17:33:00 efried jaypipes Okay, so when I remove get_available_resource, those will just vanish and nobody will care? (Assuming I'm correct in my assessment that nobody's using them today?)
17:33:40 efried It's possible those guys had their origin by us looking at the other drivers and trying to emulate them.
17:33:50 efried ...blindly, without understanding wtf we were doing :)
17:33:57 efried But that's before my time.
17:34:56 jaypipes efried: no, the idea (eventually) is to have a tree of resource providers be passed to the virt driver. the virt driver will be able to set inventory on those providers in the tree as well as decorate the providers with traits.
17:35:35 jaypipes efried: the get_inventory() patches are kind of a stepping stone towards that eventual nirvana...
17:36:22 efried jaypipes And a "resource provider" is e.g. the host, and a "resource class" is e.g. VCPU.
17:37:08 jaypipes efried: my latest thinking on this is that the resource tracker would load up a ProviderTree (from https://review.openstack.org/#/c/415920/), pass that to a virt driver update_inventory() method and then pass the tree to a virt driver set_traits() method. possibly we could do the entire update inventory and set traits in a single method...
17:37:23 jaypipes efried: yes (to answer your question above)
17:37:45 efried jaypipes Are we achieving this nirvana in pike?
17:38:00 jaypipes efried: depends on how much progress we make in merging stuff this week :)
17:38:54 efried jaypipes Okay; just trying to get a sense of how urgently I should be making my driver keep pace with this stuff.
17:39:16 jaypipes efried: not urgent.
17:39:41 jaypipes efried: the resource tracker falls back to use get_available_resource(), so don't need to worry about being left behind just yet
17:40:02 efried jaypipes From what I could tell, it only falls back if get_inventory is NotImplemented.
17:40:28 jaypipes efried: correct.
17:40:51 efried jaypipes Soooo... is it a better idea to avoid implementing get_inventory at all until more of the infra is baked?
17:41:05 jaypipes efried: yes :)
17:41:32 jaypipes efried: I was gently trying to suggest that without you wanting to hit me because you've already pushed a patch :)
17:41:46 efried jaypipes Gotcha. I'll just put a big ol' -W on that sucker, then.
17:41:52 jaypipes heh, k
17:42:21 openstackgerrit Merged openstack/nova master: Remove unused migrate_data kwarg from virt driver destroy() method https://review.openstack.org/452958
17:42:23 efried jaypipes Nah, publishing the patch was really a pretty minor part of this effort; the main thing was getting an understanding of how all of this is supposed to work. Which I was gonna need to do at some point anyawy.
17:42:26 efried anyway*
17:42:48 efried jaypipes Understanding the time frame helps a bunch, so thanks for that.
17:42:52 jaypipes np :)
17:43:21 jaypipes edleafe, bauzas, cdent, mriedem: reviews appreciated on https://review.openstack.org/#/c/462769/. thanks much in advance!
17:43:51 cdent jaypipes: in the queue
17:46:10 cdent jaypipes: when nirvana happens will the virt driver be authoritative for things in the ProviderTree, or will stuff like PCI devices still sometimes come from places like nova.conf? I ask because I was somewhat surprised to read "passed to the virt driver", instead of "from"
17:53:52 jaypipes cdent: good question. I'm torn.
17:58:22 melwitt first quotas patch is approved \o/ hallelujah
18:03:30 mriedem melwitt: did i persuade jaypipes otherwise?
18:03:41 jaypipes mriedem: yes
18:04:01 mriedem melwitt: i think https://review.openstack.org/#/c/446239/12/nova/quota.py@172 is valid though
18:04:07 mriedem but could be a separate patch
18:05:18 mriedem smcginnis and i have been trying to book flights to china all day, it's not going well
18:05:31 mriedem what was $2100 became $7K within an hour
18:05:41 smcginnis more like minutes.
18:07:04 melwitt mriedem: ah, right. that was a remnant from when I was considering putting the backward compat quota_usage_get_all* calls in _get_usages() or something like that
18:08:02 melwitt quota_usage_get_all* puts 'project_id' and 'user_id' in the dict it returns
18:08:25 mriedem yeah i was looking for that but you must have removed it
18:08:33 mriedem anyway, i'd just cleanup that docstring in a follow up change
18:09:31 melwitt which makes me wonder if I should mirror that to keep the contents similar. I didn't think about it because nothing uses the 'project_id' and 'user_id' keys
18:12:22 mriedem melwitt: if the caller doesn't care i don't see why we would
18:12:36 mriedem melwitt: btw, https://review.openstack.org/#/c/468559/ fixes a regression in cells v2 since newton, would be nice to get that backported
18:12:48 mriedem if we're quick about it i could even hold up the newton/ocata releases i requested last week
18:13:03 melwitt k, looking
18:13:04 mriedem oh nvm those were approved today
18:13:29 mriedem https://review.openstack.org/#/c/468549/ was the regression test patch btw
18:14:54 clarkb mriedem: supposedly they track your browser and such for things like that
18:15:07 clarkb mriedem: so uh maybe try private browsing and use linux
18:15:22 clarkb (supposedly OSes you pay for imply they can charge you higher rates)
18:18:34 melwitt mriedem: nice job on the func test. how did you make the servers go to ERROR state?
18:19:51 melwitt oh, nvm. just saw the comment about no compute service. derp
18:20:09 mriedem clarkb: didn't think about that
18:24:19 mriedem too late now, already booked
18:24:24 mriedem will have to remember that for next time though
18:29:58 cr0wrx anyone here familiar with any way to pull back memory dumps from instances? I don't believe it's currently a feature in nova but wanted to double check (or possibly someone knows of another project that might be aiming to do this)
19:01:34 jaypipes mriedem: question for you... I've now recheck'd this patch (https://review.openstack.org/#/c/462770/) a few times. keep getting failures, but I'm pretty sure the failures are completely unrelated to my patch. would you mind taking a look and verifying that it's not my patch that is causing these failures? :)
19:01:50 mriedem sure
19:02:50 mriedem jaypipes: looks like it is just your patch http://logs.openstack.org/70/462770/7/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/05083f7/logs/devstacklog.txt.gz#_2017-05-30_18_14_47_542
19:02:53 mriedem devstack setup fails
19:03:28 jaypipes mriedem: it *is* my patch?
19:03:44 mriedem yeah
19:03:56 mriedem https://review.openstack.org/#/c/462770/7/nova/objects/resource_provider.py@31
19:04:36 mriedem well,
19:04:41 jaypipes :( edleafe, turns out the auto-module-import thing ain't working..
19:04:42 mriedem it's something in os-traits importing things from django?
19:04:45 mriedem right
19:04:49 mriedem os-traits is fubar'ed
19:05:10 edleafe jaypipes: it's not? Since when?
19:05:21 mriedem edleafe: see the stracktrace in ^
19:05:25 mriedem *stacktrace
19:07:08 sean-k-mooney mriedem: not directly looking at the os-traits code i dont see any reference to dango
19:07:10 openstackgerrit Merged openstack/nova master: libvirt: mark some Image backend methods as abstract https://review.openstack.org/448142
19:07:35 edleafe could the __name__ be messed up somehow in the tests?
19:07:43 sean-k-mooney mriedem: its form this import https://github.com/openstack/os-traits/blob/b169eb4a8e13c74673b54512d0577088aa3cc81d/os_traits/__init__.py#L16
19:08:12 mriedem for loader, name, is_pkg in pkgutil.walk_packages(package.__path__):
19:08:30 sean-k-mooney mriedem: yes on line 52 https://github.com/openstack/os-traits/blob/b169eb4a8e13c74673b54512d0577088aa3cc81d/os_traits/__init__.py#L52

Earlier   Later