Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2019-05-07
13:55:39 mriedem huh, auto_disk_config is a valid filter parameter for listing servers but not handled in the DB API at all from what i can tell
13:55:48 mriedem oh well, it's xen only anyway
14:02:09 openstackgerrit Dan Smith proposed openstack/nova master: Add a workaround config toggle to refuse ceph image upload https://review.opendev.org/657078
14:08:38 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/stein: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655678
14:17:43 mriedem tssurya: done
14:17:46 mriedem might be a record
14:18:42 tssurya mriedem: thanks! record for a hard rain of comments?
14:18:45 tssurya :D
14:18:50 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/rocky: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655692
14:19:14 mriedem tssurya: yeah
14:19:36 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/queens: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655694
14:19:54 tssurya mriedem: heh but thanks a lot for the time, I'll work on them
14:20:40 tssurya mriedem: btw am I causing these tests to fail ? http://logs.openstack.org/62/648662/8/check/openstack-tox-py36/7604398/testr_results.html.gz
14:20:48 tssurya I couldn't find the cause
14:21:27 mriedem no
14:21:36 mriedem http://status.openstack.org/elastic-recheck/#1823251
14:21:54 tssurya mriedem: oh cool thanks
14:29:08 mriedem dansmith: did you see my question/concern on the image type request filter patch regarding bfv instances?
14:29:13 mriedem i didn't -1 yet
14:29:38 dansmith no
14:30:59 mrch_ had anyone problems upping novacompute after upgrade queens>rocky (centos)?
14:37:42 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/rocky: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655692
14:38:19 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/queens: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655694
14:49:40 openstackgerrit Matt Riedemann proposed openstack/nova master: api-ref: fix mention of all_tenants filter for non-admins https://review.opendev.org/657620
14:50:54 mriedem sorrison: you know how you added the "os_compute_api:servers:allow_all_filters" policy rule? if using all_tenants as the filter parameter, that could still be filtered out for non-admins because all_tenants has it's own specific rules (os_compute_api:servers:index:get_all_tenants and os_compute_api:servers:detail:get_all_tenants)
14:51:20 mriedem it seems that os_compute_api:servers:allow_all_filters should also apply to the all_tenants filter, yeah?
15:01:05 openstackgerrit Dan Smith proposed openstack/nova master: Expose Hyper-V supported image types https://review.opendev.org/655137
15:01:05 openstackgerrit Dan Smith proposed openstack/nova master: Make libvirt expose supported image types https://review.opendev.org/653454
15:01:06 openstackgerrit Dan Smith proposed openstack/nova master: Add ironic driver image type capabilities https://review.opendev.org/655729
15:01:06 openstackgerrit Dan Smith proposed openstack/nova master: Add vmware driver image type capabilities https://review.opendev.org/655730
15:01:07 openstackgerrit Dan Smith proposed openstack/nova master: Add xenapi driver image type capabilities https://review.opendev.org/655731
15:01:07 openstackgerrit Dan Smith proposed openstack/nova master: Add zvm driver image type capabilities https://review.opendev.org/655732
15:01:08 openstackgerrit Dan Smith proposed openstack/nova master: Add image type request filter https://review.opendev.org/656413
15:01:08 openstackgerrit Dan Smith proposed openstack/nova master: Enable image type query support in nova-next https://review.opendev.org/656903
15:01:09 openstackgerrit Dan Smith proposed openstack/nova master: Add docs for image type support request filter https://review.opendev.org/657025
15:03:10 dansmith mriedem: I've been thinking about the tracing thing too
15:05:14 dansmith more about making the existing filters log what they're doing, but the timer thing is legit too
15:08:42 mriedem i'd think that osprofiler should be handling the time spent in each filter, but not sure how granular that is
15:10:42 dansmith meaning we need some sort of hook thing?
15:12:06 mriedem i'm not sure if the @profiler.trace_cls decorator on the SchedulerAPI (rpc) gives us what we'd want there
15:12:15 mriedem since select_destinations is a call it might, but not sure
15:12:41 openstackgerrit Surya Seetharaman proposed openstack/nova master: [Docs] Change the server query parameter display into a list docs. https://review.opendev.org/657624
15:14:42 mriedem there was a recent ML thread about nicer osprofiler visualizations but i can't find it
15:16:24 mriedem oh right, something like this http://logs.openstack.org/69/617269/3/check/tempest-smoke-py3-osprofiler-redis/d5563c0/osprofiler-traces/trace-58ed6da6-f2c8-418c-9c60-337bf00ad86a.html.gz
15:16:29 openstackgerrit Surya Seetharaman proposed openstack/nova master: [Docs] Change the server query parameter display into a list. https://review.opendev.org/657624
15:16:33 mriedem not sure how to easily find a server create trace in there
15:21:37 dansmith mriedem: well, I'll start something and you can add to it if you want
15:22:08 dansmith it'd be really nice if we could see what the reqspec looks like before/after in a generic way, but that's probably too detailed and too fragile, vs. each one just logging what it's doing
15:22:39 mriedem i guess http://logs.openstack.org/69/617269/3/check/tempest-smoke-py3-osprofiler-redis/d5563c0/osprofiler-traces/trace-46016681-2a3d-4ca2-adf7-c4941649d753.html.gz and filter on nova-scheduler
15:23:08 mriedem doesn't give me the time per filter though
15:23:39 mriedem just says that like select_destinations took 302ms
15:29:47 mriedem ah timeutils already has something i was thinking to use
15:29:55 dansmith I was just using stopwatch
15:30:35 mriedem looks like you can just use oslo_utils.timeutils.time_it
15:31:27 dansmith yeah, but I want to not log if the thing appears disabled
15:32:14 mriedem the decorator takes an enabled kwarg
15:32:43 dansmith but we don't know until after it's run, unless we change something
15:32:58 dansmith I was just going to have them return a boolean if they're enabled/effective
15:33:05 dansmith like, if we're is_bfv=True, no point in logging 0.0
15:33:23 dansmith I could only log if above a threshold, but then you don't know if something is running or running fastly
15:33:24 mriedem ok that's more granular than i was thinking at first, like this is pretty easy:
15:33:25 mriedem http://paste.openstack.org/show/750876/
15:33:50 dansmith yeah, that works for the two simple ones, but not for my new one
15:34:09 dansmith because it has three reasons to not log
15:34:23 dansmith give me a few and I'll push up what I was thinking
15:34:40 mriedem map_az_to_placement_aggregate also has several reasons to not do anything
15:34:45 mriedem so depends on how granular you want this
15:38:49 openstackgerrit Dan Smith proposed openstack/nova master: WIP: Add extra logging to request filters https://review.opendev.org/657629
15:38:52 dansmith mriedem: ^
15:39:11 dansmith generic timer and contextual logs from each filter about what change it's making
15:39:20 dansmith which I think is what I'd want to validate or diagnose
15:50:10 mriedem i think your unit tests in https://review.opendev.org/#/c/656413/ are going to blow up
15:51:36 dansmith I thought is_bfv was usage safe so it would just return false
15:51:43 dansmith but I didn't run them
15:52:31 dansmith or maybe we set it somewhere before we get too far, but I won't hit that in the unit tests
15:52:56 mriedem conductor sets it before the scheduler for a move, and api for a new server, but your unit tests won't lazy-load the value
15:53:25 dansmith yeah
15:54:33 dansmith I need to remove the empty image from the volume case anyway
15:58:39 openstackgerrit Dan Smith proposed openstack/nova master: Enable image type query support in nova-next https://review.opendev.org/656903
15:58:40 openstackgerrit Dan Smith proposed openstack/nova master: Add docs for image type support request filter https://review.opendev.org/657025
16:24:14 openstackgerrit Dan Smith proposed openstack/nova master: WIP: Add extra logging to request filters https://review.opendev.org/657629
16:40:20 openstackgerrit John Garbutt proposed openstack/nova-specs master: Add Unified Limits Spec https://review.opendev.org/602201
17:23:36 dansmith ah crap
17:24:57 openstackgerrit Dan Smith proposed openstack/nova master: Add image type request filter https://review.opendev.org/656413
17:24:58 openstackgerrit Dan Smith proposed openstack/nova master: Enable image type query support in nova-next https://review.opendev.org/656903
17:24:58 openstackgerrit Dan Smith proposed openstack/nova master: Add docs for image type support request filter https://review.opendev.org/657025
17:24:59 openstackgerrit Dan Smith proposed openstack/nova master: WIP: Add extra logging to request filters https://review.opendev.org/657629
17:34:28 openstackgerrit Rodolfo Alonso Hernandez proposed openstack/os-vif stable/rocky: Prevent "qbr" Linux Bridge from replying to ARP messages https://review.opendev.org/655692
17:50:38 dansmith mriedem: have you been following the helm and health check stuff that wants to throw a non-nova tool on our bus to ping all our services all the time for health checks?
17:51:34 dansmith I've been arguing for going the real http-based healthcheck route to avoid extra message bus load, the upgrade concerns, and to, you know, allow health checks to work in the absence of a working message bus, which is the most likely failure in any deployment :)
17:52:49 artom <drive by> I'd argue that you'd still need to check message queue health, but that it's a separate thing and shouldn't be confused with Nova
17:53:16 artom Cuz if your message queue is down, not much is going to work :)
17:53:25 dansmith nova does this already, and, the http based check can report "hey I can't talk to rabbit"
17:53:46 dansmith the http check returns a json blob of "things I know about" and "are they working or not"
17:54:16 dansmith so, "can I talk to rabbit" "am I able to report my service status" "can I talk to libvirt right now" etc
17:55:07 dansmith doing that over the message bus adds load to the most loaded thing, and doesn't work if the bus is down,
17:55:27 dansmith but also won't work if conductor is down after a few minutes, because nova-compute is likely to have exhausted its threadpool trying to check in
17:55:33 dansmith so it's really a fragile expectation I think
18:15:32 artom dansmith, oh, we have a specific health check "API"?
18:15:32 artom TIL

Earlier   Later