Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-05-30
14:04:16 efried I've been kinda hoping others would have approved it by the time I got around to it.
14:11:44 openstackgerrit Adrian Chiris proposed openstack/nova stable/queens: Allow driver to properly unplug VIFs on destination on confirm resize https://review.opendev.org/661572
14:11:44 openstackgerrit Adrian Chiris proposed openstack/nova stable/queens: Move get_pci_mapping_for_migration to MigrationContext https://review.opendev.org/661571
14:18:45 sean-k-mooney oh ^ are backports. i was confused as i had thought it had merged on master
14:19:04 sean-k-mooney also nova team meeting is not right
14:39:50 gmann johnthetubaguy: on policy spec- base defined role mapping to check_str is enough or you want complete list of policy and their propoed check_str ?
14:47:41 mriedem adrianc_: do you care about this old bug fix? https://review.opendev.org/#/c/361438 if so please ack on it that you'll take it over otherwise i'm going to abandon it
14:47:57 mriedem starlingx probably cares as well but cfriesen isn't around to ask
14:50:14 johnthetubaguy gmann: I was thinking just the base, as examples, maybe add in keypairs too, but the rest we meet those problems as we hit them?
14:50:38 johnthetubaguy gmann: do you think my suggestion works, with changing the base rule based on that config?
14:51:12 gmann johnthetubaguy: +1 on base rule details i will add those.
14:51:55 efried mriedem: dtroyer might be able to answer starlingx care-about questions.
14:52:09 gmann johnthetubaguy: but till enforce_scope is false, scope_type will not have any effect so out old deprecated rule with old roles will work fine right ?
14:52:29 gmann i forgot to reply config things on gerrit.
14:53:00 cdent oh I had an open an missed it, but it's mostly directed at mriedem
14:53:04 mriedem f i suppose they have an irc channel
14:53:15 johnthetubaguy good question... its more than enforce_scope isn't enough, because its basically system or project for almost anything
14:53:35 cdent you remember a couple weeks ago when 10,000 allocation candidates was around 36 seconds? it's down to 6 now after "various changes"
14:53:45 cdent so that's like a
14:53:47 efried \o/
14:53:51 johnthetubaguy gmann: if we add the scope check in the rule by default all the time, boom, all current admins are busted
14:54:00 cdent #success 10000 allocation candidates got faster
14:54:02 openstackstatus cdent: Added success to Success page (https://wiki.openstack.org/wiki/Successes)
14:54:07 johnthetubaguy cdent: nice job
14:54:34 gmann johnthetubaguy: we will keep old check_str as deprecated rule , for example: https://review.opendev.org/#/c/645427/5/nova/policies/services.py
14:54:48 sean-k-mooney cdent: just faster in aggreate or was there one change in partcalar that stood out
14:54:59 gmann ^^ above patch was for granular change but it can be same for check_str change also
14:55:01 sean-k-mooney but also cool
14:55:20 cdent sean-k-mooney: i'll take the details to the placement channel
14:55:20 mriedem cdent: cool
14:59:53 gmann johnthetubaguy: i mean old rule(check_str without scope string) will keep working because we will keep them as deprecated rule. For example, current admin role (without scope thing) will be able to access the host APIs.
15:01:46 johnthetubaguy gmann: ah, so we are going to do the deprecated check_str idea... i forgot about that, carry on :)
15:03:11 gmann johnthetubaguy: yeah, i will add those detail in spec tomorrow if deprecated check_str solve all issue. as we are not going to do name change things it will be easy for us to do migration.
15:03:23 johnthetubaguy gmann: +1
15:03:38 artom efried, so I read the email, IIUC the current plan is to not have can_split, and just have some nodes expose NUMA and some nodes not expose NUMA
15:03:48 artom NUMA instances can go on the former, non-NUMA on the latter
15:03:52 artom Is that the tl'dr?
15:04:50 mriedem lyarwood: i just abandoned this swap volume related bug fix but you might want to check it out https://review.opendev.org/#/c/300616/
15:04:52 mriedem to see if it's still valid
15:04:53 adrianc_ mriedem: regarding https://review.opendev.org/#/c/361468, can be abandoned IMO.
15:05:22 mriedem adrianc_: i'm assuming you mean the nova change right?
15:05:25 mriedem https://review.opendev.org/#/c/361438
15:06:00 lyarwood mriedem: ack thanks for the pointer
15:06:02 adrianc_ yea
15:06:26 mriedem adrianc_: ok any particular reason why? is it no longer valid?
15:06:37 mriedem b/c the bug is still open
15:07:32 efried artom: yup, that's it
15:08:01 efried artom: the only reason we would have needed can_split is so we could put non-NUMA VMs onto hosts whose RPs were modeled as NUMA-aware trees.
15:08:02 adrianc_ mriedem: well, i would expect SR-IOV to be used and not a direct physical port.
15:08:24 mriedem artom: efried: the big problem with that is going to be really small deployments that can't split that way right?
15:08:36 mriedem i.e. an edge site with 1 compute node
15:08:59 mriedem essentially the virtual (shared) and dedicated physical cpu problem
15:09:03 artom efried, mriedem, also the fact that hugepages and CPU pinning currently add an implicit NUMA topology
15:09:04 mriedem on the same host i mean
15:09:29 adrianc_ mriedem: i.e direct port and not direct-physical. while the bug is still valid, im not too sure its really stepping on too many toes. (at least not form mellanox side :) )
15:09:31 artom So all those instances suddenly won't be able to go on "flat" nodes
15:09:32 efried yup, what we need to hear is whether that's going to be a show-stopping issue for anyone.
15:09:54 mriedem ok, probably need to ask dansmith re edge
15:10:04 mriedem not sure if he's back today or not
15:10:21 mriedem adrianc_: ok i'll mark as low priority then, thanks
15:10:23 mriedem *severity
15:10:42 adrianc_ mriedem: np
15:11:42 mriedem ok top 3 changes on http://status.openstack.org/reviews/#nova have been abandoned
15:11:46 mriedem now i'm done for 3 weeks
15:14:18 openstackgerrit Merged openstack/nova stable/queens: [stable-only] Delete allocations even if _confirm_resize raises (part 2) https://review.opendev.org/661349
15:14:37 adrianc_ stephenfin: any chance to get some attention on the FUP : https://review.opendev.org/#/c/659101 :)
15:20:35 openstackgerrit Lee Yarwood proposed openstack/nova master: WIP/DNM libvirt: flatten rbd volumes when unshelving an instance https://review.opendev.org/457886
15:20:39 openstackgerrit Matt Riedemann proposed openstack/nova master: Add validation for cpu_realtime_mask https://review.opendev.org/468203
15:20:52 mriedem another old ass change ^
15:20:54 mriedem cfriesen: ^
15:21:01 mriedem cfriesen: do you still want/need that change?
15:21:10 lyarwood melwitt / mdbooth ; ^ kicked the rbd unshelve bugfix again, using vm_state this time as suggested. I'll sort tests out later.
15:21:26 mriedem sean-k-mooney: did a neutron docs patch ever happen for the sriov live migration support?
15:21:40 melwitt lyarwood: ack
15:22:29 sean-k-mooney mriedem: not yet but it is still planned. e.g. i just finsihed working on a spec and its next on my list
15:22:31 cfriesen mriedem: I think the last comment there from stephenfin was that with the eventual changes for cpu resources in placement that change would be irrelevent.
15:22:52 mriedem cfriesen: ok so in 3 years to never
15:23:07 mriedem cfriesen: if we don't care about it let's just abandon it
15:23:17 cfriesen pretty much. I didn't care enough to keep pushing.
15:23:28 cfriesen was just a robustness thing
15:24:47 openstackgerrit Matt Riedemann proposed openstack/nova master: Add functional test for two-cell scheduler behaviors https://review.opendev.org/452006
15:24:50 mriedem cfriesen: ok please abandon if you're not going to update it
15:26:03 cfriesen done
15:26:47 mriedem thanks
15:34:26 openstackgerrit Matt Riedemann proposed openstack/nova master: Add functional test for two-cell scheduler behaviors https://review.opendev.org/452006
15:34:54 mriedem hella old review ^ now cleaned up and passing
15:35:06 mriedem gets it off the nova reviews list
15:53:23 melwitt efried: re: zuul slow queue time, my guess for the behavior is because of the way zuul prioritizes patches http://lists.openstack.org/pipermail/openstack-discuss/2018-December/000482.html vs the number of nova patches that are in a pipeline. the reason I think this is because recently I pushed a devstack patch and it began running on a node within 15 minutes. whereas my nova patches wait a minimum of 2 hours
15:56:35 efried melwitt: Interesting...
16:00:12 sean-k-mooney melwitt: ya that has chagned a few times we have gone back and forth
16:00:29 openstackgerrit Matt Riedemann proposed openstack/nova master: Remove zeromq from getting started with compute docs https://review.opendev.org/581926
16:00:31 mriedem super old trivial docs fix i cleaned up ^
16:01:31 sean-k-mooney i think the current apprch is intended to give smaller projectes equal change for running as nova but since we have a logner queue it takes longer for our jobs to start
16:02:06 sean-k-mooney or put another way to prevent tripplo jobs consome all for the capsity
16:25:46 sean-k-mooney efried: good spot on the pmu spec ill fix that shortly also i need to fix the typo stephen noted :)
16:26:12 efried sean-k-mooney: cool, I wasn't actually even sure if that one was a go at all, since it looked like stephenfin was pretty down on it :P
16:26:31 mriedem gmann: if you clean this up i'll fast approve it https://review.opendev.org/#/c/603969/16
16:26:46 sean-k-mooney well he was basically saying can we just always to it for realtime guest uncondtionally
16:27:01 gmann mriedem: thanks, doing it
16:27:18 sean-k-mooney i still think it good to allow a choice but if that is what peopel prefer that is fine with me

Earlier   Later