Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2020-07-17
15:18:36 gibi :)
15:20:58 sean-k-mooney gmann: its not just the api its also the conductor and schuler that need to agree on the db schemes and rpc verions
15:29:25 openstackgerrit Balazs Gibizer proposed openstack/nova-specs master: Support SRIOV interface attach and detach https://review.opendev.org/741601
15:29:50 gibi stephenfin, sean-k-mooney: ^^ fixed up the spec
15:30:46 melwitt lyarwood: commented, there are some odd things in the tests (args passed but not used) and tests that are simulating non-real situations, not sure if that was intended
15:32:21 lyarwood melwitt: ack looking
15:32:44 lyarwood sean-k-mooney: https://review.opendev.org/#/c/728057/ btw, looks like this change blows up when we log anything in ERROR
15:36:59 stephenfin gibi: +2 for me. Seems like a no-brainer if you're able to do the work
15:37:19 sean-k-mooney it does?
15:37:49 stephenfin sean-k-mooney: I thought you were talking to me there for a sec
15:37:50 sean-k-mooney it is not even specific to the error log
15:38:32 sean-k-mooney oh no lyarwood ^
15:39:19 sean-k-mooney gibi: looking now
15:40:45 gibi stephenfin: I think I can do the work. The WIP patch already works for me in devstack so only automatic testing and documentation is missing.
15:40:54 gibi stephenfin, sean-k-mooney: thanks
15:40:57 stephenfin \o/
15:41:28 lyarwood sean-k-mooney: yeah I assume py27 doesn't like the filter you provided?
15:41:34 sean-k-mooney gibi: +1
15:41:46 sean-k-mooney lyarwood: ah yes that is possible
15:41:46 gibi awesome, thanks
15:42:07 sean-k-mooney lyarwood: its a generator expression
15:42:08 lyarwood https://docs.python.org/2/library/logging.html#filter-objects
15:42:13 openstackgerrit Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to 'nova.pci' https://review.opendev.org/676209
15:42:13 openstackgerrit Stephen Finucane proposed openstack/nova master: mypy: Add nova.cmd, nova.conf, nova.console https://review.opendev.org/705657
15:42:14 openstackgerrit Stephen Finucane proposed openstack/nova master: mypy: Add type annotations to top-level modules https://review.opendev.org/705658
15:42:14 openstackgerrit Stephen Finucane proposed openstack/nova master: rpc: Rework 'get_notifier', 'wrap_exception' https://review.opendev.org/741663
15:42:15 melwitt lyarwood: fyi guh, the ceph job looks like it's been angry on stable branches
15:42:22 sean-k-mooney return not any(msg in log_record.msg for msg in messages)
15:42:22 lyarwood it needs to be in a filter method
15:42:36 lyarwood melwitt: yeah I have a few changes open to look at
15:42:53 melwitt ack
15:43:56 stephenfin sean-k-mooney, lyarwood: Looks like it needs to be a specific class in Python 2? https://docs.python.org/2/library/logging.html#filter
15:44:02 stephenfin https://docs.python.org/2/library/logging.html#filter-objects
15:44:47 lyarwood right
15:44:53 stephenfin actually idk, nothing in the docs by way of versionedchanged admonitions
15:45:09 stephenfin so if it's a change, it's not well documented
15:45:17 lyarwood between 3 and 2?
15:45:21 lyarwood ha
15:45:51 sean-k-mooney stephenfin: ah ok
15:46:03 sean-k-mooney ill go fix that so
15:46:12 stephenfin it might be worth checking what you've done *definitely* works on Python 3
15:46:28 stephenfin if it does, I'd be game to submit a PR to the Python docs
15:46:29 lyarwood stephenfin: I'm pretty sure we've logged stuff at ERROR in CI ;)
15:46:57 lyarwood stephenfin: oh but you mean actually works
15:47:04 sean-k-mooney https://docs.python.org/3.8/library/logging.html#logging.Filter
15:47:12 sean-k-mooney "You don’t actually need to subclass Filter: you can pass any instance which has a filter method with the same semantics."
15:47:23 stephenfin sean-k-mooney++ Good spot
15:47:33 stephenfin python-docs++
15:48:18 sean-k-mooney i didnt see that was py3 only
15:48:25 sean-k-mooney but its easy to fix
15:48:41 sean-k-mooney that said i have dinner/breakfast cooking so ill do it in an hour or so
15:50:36 openstackgerrit Lee Yarwood proposed openstack/nova master: libvirt: Handle VIR_ERR_DEVICE_MISSING when detaching devices https://review.opendev.org/741561
15:50:48 lyarwood ^ melwitt ; hopefully sorted now thanks again
15:52:09 melwitt lyarwood: I didn't catch you with my latest comment but it's just a thought anyway
15:53:42 lyarwood melwitt: yeah I'd like to do that in a FUP if possible while simplifying the logic
15:53:52 melwitt k
16:00:10 kashyap melwitt: Thanks for tha close reading of the tests review. I wasn't as careful
16:00:27 stephenfin neither was I :(
16:02:30 melwitt those tests are confusing 😝
16:02:43 kashyap melwitt: Yes! I bow to your eagle eyes. Sometimes, I give up in despair
16:03:02 melwitt not by fault of lyarwood but just the number of combos needed. I'm almost thinking ddt would have helped here
16:03:55 lyarwood melwitt: this needs to go back to stable/queens, did we have ddt back then?
16:04:23 stephenfin yup
16:04:46 stephenfin we've had that since Grizzly or something, afaik
16:04:50 lyarwood ah cool
16:05:09 lyarwood kk well I can respin again if it makes it easier to read and update in the future
16:07:08 melwitt lyarwood: well, if you're refactoring and making those error catchings go away, then it doesn't matter [after this] right? but I'm afraid some combinations are still missing so brace yourself. I might be wrong though, somehow looking at these makes me really confused
16:11:08 openstackgerrit Balazs Gibizer proposed openstack/nova master: Remove broken legacy zuul jobs https://review.opendev.org/741603
16:23:44 openstackgerrit Stephen Finucane proposed openstack/nova master: libvirt: Use better variable names, types for '_create_guest' https://review.opendev.org/740945
16:23:45 openstackgerrit Stephen Finucane proposed openstack/nova master: WIP: Attempt to restore some sanity to snapshot https://review.opendev.org/740336
16:30:11 dansmith gah, stephenfin can you tell me what the magic oslo config reno linky thing is?
16:30:17 dansmith my grepping is failing
16:30:43 dansmith or is that docs only?
16:30:46 stephenfin the sphinx roles? You can't use those in renos
16:31:16 stephenfin just use ``literals``
16:31:33 dansmith ack, that must be why I can't find it
16:32:13 stephenfin it's :oslo.config:option:`group.opt` for stuff in doc/source
16:33:00 dansmith yeah okay, I was only grepping the reno tree thinking I'd find examples in there
16:33:02 dansmith but makes sense
17:35:56 openstackgerrit sean mooney proposed openstack/nova stable/train: Silence amqp heartbeat warning https://review.opendev.org/728057
17:36:23 sean-k-mooney stephenfin: lyarwood that should be the python 2 compatiable way to do the filtering
17:36:58 sean-k-mooney but we will see what the ci says
17:37:02 stephenfin sean-k-mooney: Can you add a NOTE explaining what you've changed?
17:37:33 sean-k-mooney sure one sec ill do it now
17:37:37 stephenfin and normally I'd suggest a 'Changes:' section, but seeing as there's only one file it's probably not necessary :)
17:44:54 openstackgerrit sean mooney proposed openstack/nova stable/train: Silence amqp heartbeat warning https://review.opendev.org/728057
17:45:07 artom mordred, in my continuing sweep of sdk I stumbled upon https://review.opendev.org/#/c/741688/ - it's kinda touchy because we're breaking backwards compat (except not really?), otoh leaving it as is is overly restrictive too...
17:52:45 mnaser sean-k-mooney: have a look at what moving to openvswitch firewall driver did => https://docs.google.com/spreadsheets/d/14AzjZ8SaA6e-WOJ_0BDYAKVHtUIEobhyFWOZQPhN9s0/edit#gid=1769100043
17:52:49 mnaser switch was done at around 15UTC
17:54:08 sean-k-mooney it looks like all the long tails are gone
17:55:59 sean-k-mooney well maybe not from 1600 yes
17:56:55 sean-k-mooney i think there is still an improvement?
17:57:07 sean-k-mooney but from the graph alone its hare to tell
17:58:07 mnaser sean-k-mooney: yeah, it's still early but the results look much better so far
17:58:08 sean-k-mooney mnaser: i think the median and mean times has droped and the make latencies are better ? is that correct
17:58:29 mnaser yes, and also max time dropped (with noe or two odd ones)
17:58:36 sean-k-mooney mnaser: in this config os-vif does not need to use privesep at all i belive if you are also using native
17:58:55 mnaser sean-k-mooney: its using native but it still does because i still see a privsep process up
17:59:04 mnaser but i cant imagine why it's needed
17:59:08 sean-k-mooney that is proably novas

Earlier   Later