Wednesday, 2018-06-06

*** edmondsw has joined #openstack-powervm00:51
*** edmondsw has quit IRC00:55
*** edmondsw has joined #openstack-powervm02:39
*** edmondsw has quit IRC02:44
*** edmondsw has joined #openstack-powervm04:27
*** edmondsw has quit IRC04:32
*** chhagarw has joined #openstack-powervm05:27
*** mujahidali has joined #openstack-powervm05:57
*** edmondsw has joined #openstack-powervm06:15
*** edmondsw has quit IRC06:20
*** openstackgerrit has quit IRC07:19
*** edmondsw has joined #openstack-powervm08:03
*** edmondsw has quit IRC08:08
*** mujahidali has quit IRC09:07
*** mujahidali has joined #openstack-powervm09:07
*** YuYangWang has joined #openstack-powervm09:29
*** edmondsw has joined #openstack-powervm09:52
*** mujahidali has quit IRC09:53
*** mujahidali has joined #openstack-powervm09:54
*** edmondsw has quit IRC09:56
*** edmondsw has joined #openstack-powervm11:12
*** edmondsw has quit IRC11:16
*** edmondsw has joined #openstack-powervm11:20
*** mujahidali has quit IRC11:54
*** chhagarw has quit IRC11:58
*** AlexeyAbashkin has joined #openstack-powervm12:29
*** chhagarw has joined #openstack-powervm12:50
*** efried has quit IRC12:58
*** openstackgerrit has joined #openstack-powervm13:36
openstackgerritEric Berglund proposed openstack/nova-powervm master: DNM: ci check  https://review.openstack.org/32831513:36
*** Alexey_Abashkin has joined #openstack-powervm13:51
*** AlexeyAbashkin has quit IRC13:53
*** Alexey_Abashkin is now known as AlexeyAbashkin13:53
*** efried has joined #openstack-powervm14:09
*** esberglu has joined #openstack-powervm14:14
*** tjakobs has joined #openstack-powervm14:19
*** AlexeyAbashkin has quit IRC14:28
*** AlexeyAbashkin has joined #openstack-powervm14:36
openstackgerritEric Berglund proposed openstack/nova-powervm master: DNM: CI Check2  https://review.openstack.org/32831715:12
*** AlexeyAbashkin has quit IRC15:14
*** AlexeyAbashkin has joined #openstack-powervm15:16
*** smatzek has joined #openstack-powervm16:39
*** smatzek has quit IRC16:52
*** AlexeyAbashkin has quit IRC16:57
openstackgerritElod Illes proposed openstack/ceilometer-powervm stable/pike: Fix sphinx-docs job for stable branch  https://review.openstack.org/57284117:10
openstackgerritChhavi Agarwal proposed openstack/nova-powervm master: iSCSI Live Migration Support  https://review.openstack.org/56757517:30
*** edmondsw has quit IRC17:32
*** chhagarw has quit IRC17:37
*** edmondsw has joined #openstack-powervm17:45
openstackgerritMatthew Edmonds proposed openstack/nova-powervm master: Remove unused volume_adapter conf  https://review.openstack.org/57299520:28
edmondswefried got a question from chhagarw today that prompted me to notice this ^20:28
edmondswI guess I really need to start with "I" when I say something like that... sounds like you got the question :)20:29
efried+2, good cleanup.20:30
efriededmondsw: btw, if you're interested in what came out of the "nrp upgrade shitstorm": https://review.openstack.org/#/c/572583/20:31
efriedTL;DR: a whole bunch of extra work in Rocky that we weren't expecting.20:31
edmondswefried so it's safe to assume that any consumer of a shared provider will also be a consumer of a local provider?20:37
efriedfor now20:37
efriededmondsw: This isn't the only thing that doesn't support that case at the moment.20:37
edmondswoh, I see the note now :)20:37
edmondswcan you change the argument order of update_provider_tree to add allocations in the middle?20:39
edmondswI would have thought it needed to go on the end20:39
efriededmondsw: I asked that question early on in my nova career.  Basically, none of this is API, so we're under no obligation to preserve any kind of compatibility.20:40
efriedWe update everything in tree to conform; and we "handle" OOT drivers by sending a note to the ML.20:40
edmondswefried why isn't it a kwarg with None as the default? Most of the time you call update_provider_tree you wouldn't want to touch allocations, no?20:42
*** esberglu_ has joined #openstack-powervm20:42
efriededmondsw: The caller can never know when this shenanigans is going to happen, so he always needs to send the allocations in.20:42
edmondswthat'd be a useful note in the spec20:43
efriededmondsw: Perhaps it needs to be made clear that we expect this to be used live the vast majority of the time.  The offline FFU thing will be very rare.20:43
efriededmondsw: Cool, I'm way too forest-for-the-trees on this, please do leave a comment where you think that could be useful.20:44
efriededmondsw: btw, are you looking at source or rendered?20:44
*** esberglu has quit IRC20:44
edmondswrendered20:45
edmondswI saw that sequence thing and went "gah!"20:45
edmondswand switched to rendered20:46
edmondswefried remind me who calls update_provider_tree when?20:47
efriedThe resource tracker calls it on every periodic task interval to update_available_resources20:47
efriedsec...20:47
efriededmondsw: https://github.com/openstack/nova/blob/07652b8babc9df7e0771f5f8d8ea1d977003c129/nova/compute/resource_tracker.py#L852-L89020:48
edmondswefried doesn't look like that was already getting allocation data, so there's probably an additional perf impact to query that. Significant?20:50
efriedYes, good point.  Please comment.20:51
edmondswwill do20:51
efriedshit, that one could be significant, unless we start caching allocations.  Boo.20:51
*** esberglu_ has quit IRC20:56
*** edmondsw has quit IRC21:09
*** edmondsw has joined #openstack-powervm22:06
*** edmondsw has quit IRC22:10
*** tjakobs has quit IRC22:38
*** edmondsw has joined #openstack-powervm23:54
*** edmondsw has quit IRC23:59

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!