Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2018-02-05
16:03:50 cdent If you're just going with "not loaded" and you get something weird, isn't that you're own fault?
16:03:50 edleafe efried: https://github.com/openstack/nova/blob/master/nova/scheduler/filter_scheduler.py#L411-L422
16:04:13 cdent sigh: your
16:04:18 efried cdent: I mean, we can document it that way, but it's not very satisfying to the consumer.
16:04:30 efried cdent: And I don't agree that you necessarily want anti-affinity.
16:05:02 efried If host A is 10% full and host B is 90% full, I totally want my instances (each of which consumes 10%) to both go to host A.
16:05:02 edleafe efried: and also https://github.com/openstack/nova/blob/master/nova/scheduler/filter_scheduler.py#L426-L436
16:05:24 edleafe There is plenty of race avoidance in the code.
16:05:49 cdent I dunno, I think your host should just go where it will get what it needs, unless you tell it something special.
16:06:03 cdent And providing too many knobs is bad for everyone
16:07:08 efried edleafe: Okay, glad those mechanisms are there, but they're somewhat inelegant. (Not that thrashing GET /allocation_candidates is any more elegant - overall I think it's clunkier and less efficient - but as an example, 'twill serve.)
16:07:23 mriedem dansmith: i'd have to take a look at that flag to see what the api does with it
16:07:40 efried cdent: Too many knobs, sure, but I thought a "least consumed" weigher was one of the basics.
16:07:43 dansmith mriedem: okay I thought you did it which is why I was asking.. I'll add a note and come back to it
16:07:56 openstackgerrit Merged openstack/nova stable/pike: doc: Add user index page https://review.openstack.org/540494
16:08:02 openstackgerrit Merged openstack/nova stable/pike: Migrate "launch instance" user guide docs https://review.openstack.org/540495
16:09:06 cdent efried: is it? to me it sounds like a "exposing too much info to the user about this cloud" and "our resource accounting and allocation ratio subscription model is actually a cheat, here we've provided you with this way to get around it rather than being good cloud providers with good ratios"
16:09:13 mriedem dansmith: no i don't think so, sounds like a tdurakov thing
16:09:36 dansmith mriedem: ack
16:10:40 efried cdent: I don't see what this has to do with allocation ratios.
16:11:16 cdent if allocation ratios are set fairly and correctly, then where your instance goes shouldn't matter: it should always have good access to resources whatever host it lands one
16:11:43 cdent if hosts are oversiubscribed then of course you're going to try to do tricks to get on an underutilized host
16:12:39 efried cdent: That makes a lot of assumptions about the possible reasons for a "least consumed" deployment strategy.
16:13:05 cdent what other reason can you want other than needing to ensure you "get enough"?
16:13:53 efried cdent, edleafe: Anyway, y'all have convinced me we don't need/want this for Nova. So until/unless we get a request from some other placement consumer (which isn't quite so likely to happen as long as placement is part of nova) I can drop it.
16:18:04 openstackgerrit Balazs Gibizer proposed openstack/nova master: Add regression test for bug 1735407 https://review.openstack.org/526095
16:18:06 openstack bug 1735407 in OpenStack Compute (nova) "[Nova] Evacuation doesn't respect anti-affinity rules" [Medium,In progress] https://launchpad.net/bugs/1735407 - Assigned to Balazs Gibizer (balazs-gibizer)
16:18:56 gibi mriedem: ^^ fixed the functional test, now start fixing the bugfix patch
16:26:15 mriedem dansmith: looks like gibi made onSharedStorage optional in evacuate
16:26:36 dansmith mriedem: right, but it was required in earlier api versions
16:26:37 openstackgerrit Jay Pipes proposed openstack/nova master: only increment disk address unit for scsi devices https://review.openstack.org/538310
16:26:51 mriedem dansmith: yeah, so it looks like we could kill it from compute...
16:26:52 dansmith so I'm wondering if we have to keep the yes/no/unspecified support in the lower layers
16:27:19 jaypipes melwitt, mdbooth: really need some assistance from either/both of you on https://review.openstack.org/538310. Please see my desperate cry for help in the second unit test there...
16:27:20 mriedem unless there is some scenario where operators actually need to rely on it b/c the auto-check using the driver is wrong,
16:27:23 mriedem but that would seem like a bug
16:27:49 dansmith mriedem: well, the point being the api used to advertise that it will check your work
16:27:59 dansmith so if people used that, and we silently ignore it in compute now...
16:28:22 mriedem it used it for some other stuff too, related to the adminPass
16:29:23 mriedem reading the spec https://specs.openstack.org/openstack/nova-specs/specs/mitaka/implemented/optional-shared-storage-flag-in-evacuate-api.html
16:30:02 openstackgerrit Eric Fried proposed openstack/nova master: placement doc: Conflict caveat for DELETE APIs https://review.openstack.org/540919
16:30:16 efried cdent, edleafe: ^ For the remainder of bug 1746373
16:30:17 openstack bug 1746373 in OpenStack Compute (nova) "Placement APIs with missing conflict detection" [Undecided,In progress] https://launchpad.net/bugs/1746373 - Assigned to Eric Fried (efried)
16:30:30 efried jaypipes also ^
16:32:03 mriedem dansmith: fine to just leave it in for compat
16:39:15 openstackgerrit Matt Riedemann proposed openstack/nova master: Rename the 'recreate' param to rebuild_instance to 'evacuate' https://review.openstack.org/508190
16:39:54 openstackgerrit Matt Riedemann proposed openstack/nova master: Rename the 'recreate' param in rebuild_instance to 'evacuate' https://review.openstack.org/508190
16:47:42 hrw stephenfin: what is docstring?
16:47:52 openstackgerrit Merged openstack/python-novaclient master: nova limits ERROR (Exception): Field names must be unique https://review.openstack.org/527922
16:47:53 openstackgerrit Merged openstack/python-novaclient master: Implement hypervisor hostname exact pattern match https://review.openstack.org/520187
16:49:38 hrw ah. """ D:
16:50:36 dansmith mriedem: ack
16:52:40 mriedem artom: https://review.openstack.org/#/c/508190/2/nova/compute/manager.py@2663
16:53:13 artom mriedem, ah, doh
16:53:28 artom mriedem, well, my second one still kinda stands :P
16:53:33 openstackgerrit Marcin Juszkiewicz proposed openstack/nova master: Make sure that we have usable input for graphical console https://review.openstack.org/538003
16:53:37 hrw stephenfin: as you wish ;D
16:53:56 openstackgerrit Merged openstack/python-novaclient master: Add CLI to show instance usage audit logs https://review.openstack.org/409413
16:54:02 artom I'll remove the -1 though
16:57:01 stephenfin hrw: ++
16:59:29 hrw mriedem, jaypipes, sahid: can you take another look at https://review.openstack.org/538003 patch? it is simplest of all versions and comes with extra test
17:01:06 stephenfin bauzas: Gentle reminder for https://review.openstack.org/#/c/530743/
17:03:26 mriedem stephenfin: see the latest comment there, it introduces a bug
17:04:04 mriedem so a big -1 from me on the backport
17:04:30 jaypipes hrw: I never saw a response from you to sahid's review comments?
17:05:04 hrw jaypipes: rewrote whole as an answer
17:05:33 stephenfin mriedem: Yup, I was focussing on the latter part of the comment though: a partial fix was better than no fix
17:05:37 jaypipes hrw: heh, ok :)
17:05:56 jaypipes hrw: I don't see a followup from sahid which is why I asked..
17:06:03 stephenfin Probably best to wait for the fixes' fix to be merged though
17:06:06 hrw jaypipes: sure
17:07:17 hrw jaypipes: when 'a reply' would be just message then I write it. If it is simple fix then I mark Done'
17:07:43 hrw jaypipes: if it is complicated then do code mostly
17:08:09 hrw jaypipes: and here comment was more or less 'rewrite the shit' so I did
17:11:57 jaypipes hrw: understood. I'd still like to have sahid's opinion on it, though, since he had the biggest objection to it..
17:12:04 hrw sure
17:28:50 hrw I have an idea for another 'lets add some device' patch just not found yet where to plug it
17:29:36 hrw on pcie machines (aarch64, x86/q35) more pcie-root-port controllers can be handy if someone wants to add more disks/networks
17:34:35 openstackgerrit Matt Riedemann proposed openstack/nova master: Add nova-status check for ironic flavor migration https://review.openstack.org/527541
17:34:39 openstackgerrit Eric Fried proposed openstack/nova master: placement doc: Conflict caveat for DELETE APIs https://review.openstack.org/540919
17:37:47 hrw it is like: basic VM instance config gets created with (let's say) 4 pcie root ports (think of them as pcie slot on virtual motherboard). then we plug video card, usb host controller (and usb keyboard to it) so we have just 2 slots left. then we add RNG and memory balloon... hm. have to boot an instance and check
17:44:19 openstackgerrit Merged openstack/nova master: Update cells v2 layout doc caveats for Queens https://review.openstack.org/540667
17:44:29 openstackgerrit Merged openstack/nova stable/ocata: libvirt: Re-initialise volumes, encryptors, and vifs on hard reboot https://review.openstack.org/531422
17:49:28 hrw arm64 guest has 1 pcie slot free when booted with graphical console
17:59:54 openstackgerrit Balazs Gibizer proposed openstack/nova master: Add late server group policy check to rebuild https://review.openstack.org/525242
18:00:49 gibi mriedem: fixed the bugfix patch as well ^^
18:01:42 mriedem ok
18:01:48 mriedem will take a look after lunch
18:02:57 gibi mriedem: have nice lunch
18:03:50 gibi mriedem: I'm no checking your idea to use wait_for_versioned_notification instead of the for loop and the sleep int he functional test, but I have to leave soon so cannot promise I finish with that today
18:04:17 mriedem that's fine, it can be a later cleanup
18:05:03 gibi mriedem: OK
18:08:53 mriedem cfriesen: i have an idea about the perf hit in pike,
18:09:04 mriedem maybe versioned notifications?
18:09:09 mriedem cfriesen: are you running with those on?
18:09:46 mriedem default is to send both versioned and unversioned notifications,
18:10:00 mriedem and the versioned notifications have to sometimes lazy-load fields on the instance for the notification payload
18:10:05 mriedem which means more round trips over rpc to the db
18:12:24 jaypipes hrw: reviewed.
18:15:00 cfriesen mriedem: will check

Earlier   Later