Thursday, 2019-08-15

openstackgerritEric Fried proposed openstack/placement master: Add tests demonstrating overlapping same_subtreeZ  https://review.opendev.org/67653000:03
*** takashin has joined #openstack-placement01:35
*** tetsuro has joined #openstack-placement02:31
*** tetsuro has quit IRC02:32
*** tetsuro has joined #openstack-placement02:37
*** tetsuro has quit IRC02:41
*** tetsuro has joined #openstack-placement02:42
*** tetsuro has quit IRC02:44
*** takashin has left #openstack-placement03:01
*** altlogbot_2 has quit IRC03:16
*** altlogbot_0 has joined #openstack-placement03:18
*** tetsuro has joined #openstack-placement03:45
*** tetsuro has quit IRC03:48
*** belmoreira has joined #openstack-placement06:55
*** tssurya has joined #openstack-placement07:32
*** helenafm has joined #openstack-placement07:42
*** e0ne has joined #openstack-placement08:15
*** gibi has quit IRC08:42
*** cdent has joined #openstack-placement08:52
*** gibi has joined #openstack-placement08:59
*** gibi has quit IRC09:05
*** gibi has joined #openstack-placement09:06
openstackgerritChris Dent proposed openstack/placement master: Add gabbits using a DeepNUMANetworkFixture  https://review.opendev.org/67620409:27
openstackgerritChris Dent proposed openstack/placement master: Add rw_ctx.parent_uuid_by_rp_uuid, for clarity  https://review.opendev.org/67623309:27
openstackgerritChris Dent proposed openstack/placement master: Use rp.id instead of uuid in _rp_rc_key  https://review.opendev.org/67623409:27
openstackgerritChris Dent proposed openstack/placement master: Add a rw_ctx.psum_res_by_rp_rc, for clarity  https://review.opendev.org/67623509:27
openstackgerritChris Dent proposed openstack/placement master: Avoid duplicate ProviderSummary in _merge_candidates  https://review.opendev.org/67560609:27
cdentgibi: overnight Eric did a bunch of investigation to figure some things out. I've rebased my stack ^ onto a test he added09:27
cdentwe're both pretty satisfied that we understand what's going on now, but it required some learning on what same_subtree really means, which we'll need to make sure gets reflected in docs before the end of the cycle09:28
cdentstephenfin: you might enjoy that stack too. if you've managed to avoid thinking "damn, placement does some magical weird", this will kill it09:29
gibicdent: ack, I will dig into that series09:30
openstackgerritChris Dent proposed openstack/placement master: Add gabbits using a DeepNUMANetworkFixture  https://review.opendev.org/67620409:53
openstackgerritChris Dent proposed openstack/placement master: Add rw_ctx.parent_uuid_by_rp_uuid, for clarity  https://review.opendev.org/67623309:53
openstackgerritChris Dent proposed openstack/placement master: Use rp.id instead of uuid in _rp_rc_key  https://review.opendev.org/67623409:53
openstackgerritChris Dent proposed openstack/placement master: Add a rw_ctx.psum_res_by_rp_rc, for clarity  https://review.opendev.org/67623509:53
openstackgerritChris Dent proposed openstack/placement master: Avoid duplicate ProviderSummary in _merge_candidates  https://review.opendev.org/67560609:53
cdentgibi: I fixed a nit from eric (who apparently never sleeps)09:53
*** belmoreira has quit IRC10:03
*** helenafm has quit IRC10:04
*** helenafm has joined #openstack-placement10:15
gibicdent: ack. I haven't started reading yet10:31
*** belmoreira has joined #openstack-placement11:09
*** belmoreira has quit IRC11:11
*** belmoreira has joined #openstack-placement11:14
openstackgerritMerged openstack/placement master: Fix allocation bug in NUMANetworkFixture  https://review.opendev.org/67648611:23
*** helenafm has quit IRC13:02
*** tssurya has quit IRC13:03
*** tssurya has joined #openstack-placement13:04
*** mriedem has joined #openstack-placement13:18
*** belmoreira has quit IRC13:45
*** belmoreira has joined #openstack-placement13:50
*** artom has joined #openstack-placement15:18
openstackgerritChris Dent proposed openstack/placement master: WIP: Use orjson in python3 for allocation candidate dump  https://review.opendev.org/67466115:26
*** belmoreira has quit IRC15:26
*** belmoreira has joined #openstack-placement15:28
efriedcdent: what's the conclusion on consumer types?15:29
cdentefried: nobody is feeling super urgent about it. It would be "nice" to have it ready by U so nova could get started with it instantly but not critical15:30
efriedack15:30
cdenttssurya has more pressing matters and nobody else has yet stepped up15:30
efriedanything else you'd like to report by proxy in today's meeting?15:30
cdentdepending on my disposition I might mess with it15:30
efriedOkay, I'll save my edit and leave you to it.15:31
efriedoh, you meant consumer types, not the meeting agenda.15:31
cdentah, the meeting15:32
cdent"we did more performance stuff"15:32
efriedyeah15:32
cdent"we discovered same_subtree needs some good usage docs so people dont make the wrong intuitions"15:32
cdent"osc_placement needs more attention"15:33
cdentEOF15:33
efriedthanks15:34
*** belmoreira has quit IRC15:38
*** efried has quit IRC15:46
cdentsigh. I just can't seem to find a way to make build_provider_summaries fast: when we need 9000 of them, we need 9000 of them...15:50
*** efried has joined #openstack-placement15:50
* cdent comes back later15:53
*** cdent has quit IRC15:53
*** tssurya has quit IRC16:18
*** e0ne has quit IRC16:28
*** cdent has joined #openstack-placement17:10
*** cdent has quit IRC17:19
*** e0ne has joined #openstack-placement17:52
*** e0ne has quit IRC18:36
*** spatel has joined #openstack-placement20:48
*** takashin has joined #openstack-placement20:48
*** spatel has quit IRC20:53
*** mriedem has quit IRC21:00
*** artom has quit IRC21:30
*** spatel has joined #openstack-placement23:48
*** spatel has quit IRC23:53

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