| *** Sundar has quit IRC | 01:12 | |
| *** takashin has joined #openstack-placement | 01:19 | |
| *** zzzeek has quit IRC | 05:07 | |
| *** zzzeek has joined #openstack-placement | 05:08 | |
| *** e0ne has joined #openstack-placement | 05:44 | |
| *** e0ne has quit IRC | 06:26 | |
| *** e0ne has joined #openstack-placement | 06:36 | |
| *** e0ne has quit IRC | 07:13 | |
| *** helenafm has joined #openstack-placement | 07:27 | |
| *** tssurya has joined #openstack-placement | 07:55 | |
| *** takashin has left #openstack-placement | 08:00 | |
| openstackgerrit | Chris Dent proposed openstack/placement master: WIP: Nested performance testing https://review.opendev.org/665695 | 08:12 |
|---|---|---|
| *** tetsuro has joined #openstack-placement | 08:14 | |
| *** ttsiouts has joined #openstack-placement | 08:16 | |
| *** e0ne has joined #openstack-placement | 08:23 | |
| *** ttsiouts has quit IRC | 08:29 | |
| *** ttsiouts has joined #openstack-placement | 08:29 | |
| *** ttsiouts_ has joined #openstack-placement | 08:30 | |
| *** ttsiouts has quit IRC | 08:30 | |
| *** dklyle has quit IRC | 08:35 | |
| *** david-lyle has joined #openstack-placement | 08:35 | |
| *** helenafm has quit IRC | 08:57 | |
| *** cdent has joined #openstack-placement | 09:23 | |
| *** ttsiouts_ has quit IRC | 10:23 | |
| *** ttsiouts has joined #openstack-placement | 10:24 | |
| *** ttsiouts has quit IRC | 10:28 | |
| *** helenafm has joined #openstack-placement | 10:33 | |
| *** tetsuro has quit IRC | 10:56 | |
| openstackgerrit | Chris Dent proposed openstack/placement master: WIP: Nested performance testing https://review.opendev.org/665695 | 10:59 |
| cdent | sigh, the turnaround loop on this is tedious | 11:00 |
| *** artom has quit IRC | 12:22 | |
| *** artom has joined #openstack-placement | 12:23 | |
| *** artom has quit IRC | 12:33 | |
| *** artom has joined #openstack-placement | 12:33 | |
| *** bauzas has left #openstack-placement | 13:04 | |
| *** bauzas has joined #openstack-placement | 13:04 | |
| *** mriedem has joined #openstack-placement | 13:06 | |
| *** ttsiouts has joined #openstack-placement | 13:06 | |
| bauzas | gibi: ++ for the placement graphtool | 13:31 |
| bauzas | thanks | 13:31 |
| *** ttsiouts has quit IRC | 13:32 | |
| gibi | bauzas: I needed something to visualize complex trees | 13:32 |
| *** ttsiouts has joined #openstack-placement | 13:33 | |
| *** ttsiouts_ has joined #openstack-placement | 13:36 | |
| *** ttsiouts has quit IRC | 13:37 | |
| cdent | gibi: have you been doing some real work with real world complex trees, or is it still mostly experimentation? | 13:38 |
| gibi | cdent: the bandwidht work already introduced enough child RPs that I needed a tool for troubleshooting | 13:38 |
| cdent | gibi: yeah, I just mean are you or people you work with actively using that stuff in the field | 13:39 |
| gibi | cdent: no active user other than me yet | 13:39 |
| efried | o/ folks | 13:45 |
| efried | cdent, gibi: I plan to update the nested magic spec today to describe root_required. | 13:46 |
| cdent | ✔ | 13:46 |
| gibi | ack | 13:47 |
| *** e0ne has quit IRC | 14:28 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: WIP: Microversion 1.35: root_required https://review.opendev.org/665492 | 14:34 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Miscellaneous doc/comment/log cleanups https://review.opendev.org/665691 | 14:34 |
| *** ttsiouts_ has quit IRC | 14:36 | |
| *** ttsiouts has joined #openstack-placement | 14:37 | |
| *** ttsiouts has quit IRC | 14:41 | |
| *** ttsiouts has joined #openstack-placement | 14:48 | |
| *** tssurya has quit IRC | 14:49 | |
| *** e0ne has joined #openstack-placement | 14:52 | |
| efried | cdent: got link to the ML discussion about 'traits flow down' at your fingertips? | 15:06 |
| efried | I found 'aggregates flow down', but not 'traits'. | 15:06 |
| cdent | one sec | 15:06 |
| cdent | in the aggregates flow down thread I point back to the middle of the thread where traits flow down came in: http://lists.openstack.org/pipermail/openstack-discuss/2019-April/005201.html | 15:07 |
| cdent | the top of which is http://lists.openstack.org/pipermail/openstack-discuss/2019-April/004817.html | 15:08 |
| efried | thanks | 15:11 |
| *** e0ne has quit IRC | 15:28 | |
| *** ttsiouts has quit IRC | 15:33 | |
| *** ttsiouts has joined #openstack-placement | 15:34 | |
| *** ttsiouts has quit IRC | 15:38 | |
| *** mgagne has quit IRC | 15:39 | |
| *** mgagne has joined #openstack-placement | 15:42 | |
| *** efried has quit IRC | 15:44 | |
| *** efried has joined #openstack-placement | 15:49 | |
| cdent | gibi, efried : I did a bit of profiling on the allocation mappings, nothing shocking, but a few thoughts, see: https://storyboard.openstack.org/#!/story/2005889 | 15:52 |
| efried | nice cdent | 15:54 |
| efried | and yeah, no big surprise. _merge_candidates is a dog. | 15:54 |
| * cdent nods | 15:55 | |
| efried | Not sure how to pare that down, really. Reducing redundant db reads will likely be easier. | 15:55 |
| *** helenafm has quit IRC | 15:56 | |
| cdent | Yeah, anything we might do in those would be fairly magical | 16:01 |
| cdent | efried: I think a next possible thing to do is turn on db debugging and inspect the real set of db queries that does one request | 16:02 |
| cdent | zzzeek: If you're about: what's your feeling on using in_() with 1000-10000 literal values. I'm looking at placement queries and we've got a few http api requests that result in a mix of queries with several of those | 16:32 |
| cdent | note to self: OS_PLACEMENT_DATABASE__CONNECTION_DEBUG=50 | 16:36 |
| *** cdent has quit IRC | 16:47 | |
| *** e0ne has joined #openstack-placement | 16:51 | |
| *** yikun has quit IRC | 16:56 | |
| zzzeek | cdent: I'd keep it closer to 1000 and use batches | 17:00 |
| zzzeek | cdent: sqlalchemy also supports a construct called "expanding" bindparam which i recommend it will become the standard in 1.4 or 2.0 | 17:00 |
| *** altlogbot_2 has quit IRC | 17:00 | |
| *** altlogbot_3 has joined #openstack-placement | 17:02 | |
| *** irclogbot_1 has quit IRC | 17:02 | |
| *** irclogbot_1 has joined #openstack-placement | 17:03 | |
| efried | zzzeek: cdent may know what you meant by "batches", but can you enlighten me? Does it simply mean splitting the list up in python and running the db query more than once, or is there some actual way in sqla to say "run this in batches where $input_list is split into batches of $size"? | 17:41 |
| *** e0ne has quit IRC | 18:13 | |
| *** e0ne has joined #openstack-placement | 18:29 | |
| *** e0ne has quit IRC | 18:33 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: Spec for nested magic 1 https://review.opendev.org/662191 | 18:35 |
| *** efried has quit IRC | 19:21 | |
| *** efried has joined #openstack-placement | 19:35 | |
| *** e0ne has joined #openstack-placement | 19:38 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: Microversion 1.35: root_required https://review.opendev.org/665492 | 20:34 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Miscellaneous doc/comment/log cleanups https://review.opendev.org/665691 | 20:34 |
| *** e0ne has quit IRC | 20:36 | |
| efried | cdent, gibi, tetsuro: Spec updated, series unWIPed. | 20:36 |
| *** mriedem is now known as mriedem_away | 21:47 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!