Friday, 2020-11-13

*** jmasud has joined #openstack-meeting00:14
*** jmasud has quit IRC00:23
*** jmasud has joined #openstack-meeting00:29
*** jmasud has quit IRC00:41
*** rcernin has quit IRC00:42
*** rcernin has joined #openstack-meeting00:43
*** jmasud has joined #openstack-meeting00:48
*** jmasud has quit IRC01:23
*** jmasud has joined #openstack-meeting01:26
*** _mlavalle_1 has quit IRC01:31
*** jmasud has quit IRC01:49
*** rcernin has quit IRC01:58
*** macz_ has quit IRC02:09
*** tinwood has quit IRC02:10
*** tinwood has joined #openstack-meeting02:13
*** rh-jelabarre has quit IRC02:33
*** macz_ has joined #openstack-meeting02:39
*** macz_ has quit IRC02:45
*** jmasud has joined #openstack-meeting02:45
*** rcernin has joined #openstack-meeting02:57
*** jmasud has quit IRC02:58
*** jmasud has joined #openstack-meeting03:05
*** jmasud has quit IRC03:09
*** rfolco has quit IRC03:14
*** jmasud has joined #openstack-meeting03:16
*** jmasud has quit IRC03:20
*** jmasud has joined #openstack-meeting03:21
*** jmasud has quit IRC03:25
*** jmasud has joined #openstack-meeting03:27
*** jmasud has quit IRC03:29
*** rcernin has quit IRC03:38
*** jmasud has joined #openstack-meeting03:39
*** rbudden has joined #openstack-meeting03:46
*** armax has quit IRC03:47
*** jmasud has quit IRC03:50
*** manpreet has quit IRC03:58
*** jmasud has joined #openstack-meeting04:03
*** rcernin has joined #openstack-meeting04:05
*** jmasud has quit IRC04:13
*** hyunsikyang__ has joined #openstack-meeting04:20
*** hyunsikyang has quit IRC04:23
*** jmasud has joined #openstack-meeting04:27
*** jmasud has quit IRC04:30
*** jmasud has joined #openstack-meeting04:33
*** jmasud has quit IRC04:35
*** jmasud has joined #openstack-meeting04:39
*** jmasud has quit IRC04:43
*** psachin has joined #openstack-meeting04:52
*** jmasud has joined #openstack-meeting04:53
*** jmasud has quit IRC05:02
*** gyee has quit IRC05:24
*** evrardjp has quit IRC05:33
*** evrardjp has joined #openstack-meeting05:33
*** rbudden has quit IRC05:33
*** baojg has quit IRC05:34
*** baojg has joined #openstack-meeting05:35
*** rcernin has quit IRC05:39
*** ociuhandu has joined #openstack-meeting05:40
*** rcernin has joined #openstack-meeting05:42
*** ociuhandu has quit IRC06:00
*** manpreet has joined #openstack-meeting06:15
*** tosky has joined #openstack-meeting06:24
*** jmasud has joined #openstack-meeting06:35
*** jmasud has quit IRC06:36
*** jmasud has joined #openstack-meeting06:37
*** rcernin_ has joined #openstack-meeting06:53
*** rcernin has quit IRC06:54
*** rcernin_ has quit IRC06:59
*** hyunsikyang has joined #openstack-meeting07:06
*** hyunsikyang__ has quit IRC07:09
*** jmasud has quit IRC07:12
*** ociuhandu has joined #openstack-meeting07:13
*** ralonsoh has joined #openstack-meeting07:14
*** ociuhandu has joined #openstack-meeting07:14
*** imcsk8 has quit IRC07:14
*** imcsk8 has joined #openstack-meeting07:16
*** ociuhandu has quit IRC07:17
*** rpittau|afk is now known as rpittau07:20
*** slaweq has joined #openstack-meeting07:26
*** david-lyle has joined #openstack-meeting08:47
*** dklyle has quit IRC08:47
*** ociuhandu has joined #openstack-meeting08:58
*** david-lyle has quit IRC08:58
*** ociuhandu has quit IRC09:08
*** jawad_axd has joined #openstack-meeting09:22
*** ociuhandu has joined #openstack-meeting09:48
*** ociuhandu has quit IRC09:58
*** baojg has quit IRC10:15
*** dsariel has quit IRC10:28
*** icey has quit IRC10:30
*** icey has joined #openstack-meeting10:36
*** jawad_axd has quit IRC11:01
*** ociuhandu has joined #openstack-meeting11:03
*** ociuhandu has quit IRC11:27
*** takamatsu has quit IRC11:33
*** takamatsu has joined #openstack-meeting11:33
*** e0ne has joined #openstack-meeting11:35
*** dsariel has joined #openstack-meeting11:36
*** ociuhandu has joined #openstack-meeting11:44
*** ykatabam has joined #openstack-meeting11:48
*** dsariel has quit IRC11:48
*** ykatabam has quit IRC11:49
*** rfolco has joined #openstack-meeting11:59
*** baojg has joined #openstack-meeting12:03
*** jmasud has joined #openstack-meeting12:23
*** jmasud has quit IRC12:36
*** rh-jelabarre has joined #openstack-meeting13:09
*** rh-jelabarre has quit IRC13:13
*** rh-jelabarre has joined #openstack-meeting13:21
*** ociuhandu_ has joined #openstack-meeting13:21
*** raildo has joined #openstack-meeting13:24
*** ociuhandu has quit IRC13:25
*** mahatic has quit IRC13:39
*** obondarev has joined #openstack-meeting13:54
*** mlavalle has joined #openstack-meeting13:58
*** baojg has quit IRC13:58
*** baojg has joined #openstack-meeting13:59
slaweq#startmeeting neutron_drivers14:00
openstackMeeting started Fri Nov 13 14:00:07 2020 UTC and is due to finish in 60 minutes.  The chair is slaweq. Information about MeetBot at http://wiki.debian.org/MeetBot.14:00
openstackUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
*** openstack changes topic to " (Meeting topic: neutron_drivers)"14:00
*** lajoskatona has joined #openstack-meeting14:00
openstackThe meeting name has been set to 'neutron_drivers'14:00
mlavalleo/14:00
lajoskatonaHi14:00
slaweqwelcome everyone!14:00
obondarevhi14:00
slaweqlets wait few more minutes for other people to join14:01
ralonsohhi14:01
slaweqping haleyb: njohnston amotoki yamamoto14:02
slaweqmaybe they will join us soon14:02
mlavalledon't pay them today14:04
obondarev:D14:04
slaweqLOL14:05
haleybhi, sorry i'm late14:05
slaweqok, lets start14:07
slaweqwe quorum already14:07
slaweqand we have only one topic in the on demand agenda for today14:07
slaweqno new RFEs to discuss14:08
slaweq#topi On Demand14:08
slaweq#topic On Demand14:08
*** openstack changes topic to "On Demand (Meeting topic: neutron_drivers)"14:08
slaweqobondarev: topic was added by You so please go on :)14:08
obondarevyep, it's about https://bugs.launchpad.net/neutron/+bug/188752314:09
openstackLaunchpad bug 1887523 in neutron "Deadlock detection code can be stale" [High,In progress]14:09
obondarevso let me briefly describe my point14:09
obondarevI think the bug with it's current title, description and suggested approach is not quite correct14:09
obondarev"Deadlock detection code can be stale" - don't think so14:10
obondarev"neutron has it's own implementation of it which is missing a bunch of deadlocks" - not True14:10
obondarevso indeed neutron has a few decorators on top of oslo.db one14:11
lajoskatonayeah perhaps "neutron has its own extra decorators around oslo decorator"14:11
obondarevbut they were added for a reason14:11
obondarevfor many reasons in fact14:11
obondarevneutron had a long history of fighting against DB errors in concurrent scenarios14:12
obondarevgit blame may shed some light on it: https://github.com/openstack/neutron/blame/stable/rocky/neutron/db/api.py14:12
obondarevjust a few examples14:12
obondarevhttps://bugs.launchpad.net/neutron/+bug/159607514:12
openstackLaunchpad bug 1596075 in neutron "Neutron confused about overlapping subnet creation" [High,Fix released] - Assigned to Kevin Benton (kevinbenton)14:12
obondarevhttps://bugs.launchpad.net/neutron/+bug/161279814:12
openstackLaunchpad bug 1612798 in neutron "Move db retry logic closer to where DB error occur" [Critical,Fix released] - Assigned to Kevin Benton (kevinbenton)14:12
obondarevhttps://github.com/openstack/neutron/commit/948461c8b2fbeb30e4fa3a43cc523cff76327d4e14:13
obondarevmost were quite a tricky ones14:13
obondarevso I don't think moving back to oslo.db decorator is the right thing to do14:14
obondarevthere were not much work done in oslo.db: https://github.com/openstack/oslo.db/commits/master/oslo_db/api.py14:14
mlavalleyou mean lately?14:14
obondarevright14:14
obondarevthanks14:14
obondarevin fact oslo.db retries were started by neutron folks)14:15
mlavalleLOL14:15
obondarevIIRC14:15
obondarevand in fact neutron retry logic now handles more cases14:15
slaweqbut from what is written in the bug description we are missing some of deadlocks and not handling them properly14:15
obondarevsome more info could be found here: https://github.com/openstack/neutron/blob/master/doc/source/contributor/internals/retries.rst14:15
obondarevslaweq, correct, so I think the bug should be about quota deadlocks14:16
slaweqmaybe we should change this LP to something like "compare our implementation with oslo db and update our where it's needed"14:16
obondarevnot about bad neutron retry logic14:16
mlavallemhhhhh....14:17
obondarevat least the bug should clearly show where neutron retries are bad14:17
*** macz_ has joined #openstack-meeting14:17
lajoskatonayeah that 's perhaps missing14:17
obondarevas if we just start replacing retry_if_session_inactive - we may got regressions14:17
mlavallein principle I might agree wioth you obondarev. However, I want to point out that Mohammed is the CEO of a big operator14:17
obondarevwjich might be hard to spot14:17
lajoskatonaif I undrstand well from summit generally there are scaling issues with neutron and one thing was that db issues can be behind that14:18
obondareveven with Loki service plugin14:18
mlavalleso at the very least I would like to get more input from mnaser and hear more about his point of view14:18
obondarevso in the bug comment and proposed patch I agree that quota retries have issues14:18
ralonsohlet me remember you that we are still migrating to the new DB engine facade14:19
ralonsohthat will remove subtransactions14:19
mlavallethat is also true14:19
obondarevbut please let's not just replace retry_if_session_inactive for oslo-db.wrap_db_retries all over neutron14:19
ralonsohand other issues related to mixing both facades14:19
ralonsohso I recommend start investigating this after completing the migration14:20
obondarevralonsoh, so there is no evidence what's exactly wrong with quota deadlock issue, right?14:20
ralonsohI can't tell, sorry14:20
ralonsohbut14:20
ralonsohwith the new facade at least we know each thread has one single transaction without subtransactions14:21
ralonsohand the context is unique14:21
ralonsoh(per thread)14:21
ralonsohe.g.:  https://review.opendev.org/#/c/715315/20/neutron/tests/functional/services/portforwarding/test_port_forwarding.py14:21
patchbotpatch 715315 - neutron - Finish the new DB engine facade migration - 20 patch sets14:21
*** macz_ has quit IRC14:22
obondarevgreat, so I'd suggest we put this bug on hold for now, and after new facade is there - try to reproduce and investigate14:22
ralonsoh+1 to this14:22
mlavalleI propose three steps:14:22
mlavalle1) finish the migration to the new db engine facade, as indicated by ralonsoh and obondarev14:23
mlavalle2) Limit this bug, for the time being, to the quota issue14:23
mlavalle3) Seek more input from mnaser. He may giove us some good insights, given his operational experience14:23
obondarev+114:24
slaweqmlavalle++14:24
lajoskatonasounds good14:24
ralonsoh+1 to this proposal14:24
slaweqand also if there will be another similar issues to what we have now with quota, lets treat them separately as regular bugs14:24
* mnaser is happy to chime in -- just ping me via email/ml (cc'd directly) or any other way :)14:24
ralonsohwe can always increase the scope of 2) to other code places14:25
mlavallethanks mnaser !14:25
njohnstonmakes sense to me14:25
slaweqok, so I think we have agreement on that for now14:26
slaweqI will sum it up in the LP comment after the meeting14:26
lajoskatonathanks slaweq14:26
obondarevthanks slaweq14:26
slaweqso I think this topic is done for today, right?14:27
obondarev+14:27
mlavalleI think so14:27
slaweqdo You have anything else You want to discuss today? if not, I will give You 30 minutes back :)14:28
mlavalleyaaay, weekened is now a bit closer!14:28
slaweqLOL14:28
slaweqfor me it's almost there :)14:28
lajoskatonait's time :-)14:28
obondarevsame here)14:28
slaweqok, thx for attending the meeting14:28
ralonsohbye! enjoy the weekend14:29
slaweqand have a great weekend!14:29
njohnstono/14:29
obondarevbye!14:29
lajoskatonaBye14:29
mlavalleo/14:29
slaweq#endmeeting14:29
*** openstack changes topic to "OpenStack Meetings || https://wiki.openstack.org/wiki/Meetings/"14:29
openstackMeeting ended Fri Nov 13 14:29:09 2020 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)14:29
openstackMinutes:        http://eavesdrop.openstack.org/meetings/neutron_drivers/2020/neutron_drivers.2020-11-13-14.00.html14:29
*** lajoskatona has left #openstack-meeting14:29
openstackMinutes (text): http://eavesdrop.openstack.org/meetings/neutron_drivers/2020/neutron_drivers.2020-11-13-14.00.txt14:29
openstackLog:            http://eavesdrop.openstack.org/meetings/neutron_drivers/2020/neutron_drivers.2020-11-13-14.00.log.html14:29
haleybo/14:29
*** dmacpher has joined #openstack-meeting14:33
*** ociuhandu_ has quit IRC14:42
*** dsariel has joined #openstack-meeting14:42
*** ociuhandu has joined #openstack-meeting14:43
*** lpetrut has joined #openstack-meeting14:49
*** rpittau is now known as rpittau|afk14:49
*** hemna has quit IRC14:50
*** hemna has joined #openstack-meeting14:51
*** ociuhandu has quit IRC15:06
*** ociuhandu has joined #openstack-meeting15:08
*** obondarev has quit IRC15:10
*** lpetrut has quit IRC15:12
*** ociuhandu has quit IRC15:24
*** ianychoi has joined #openstack-meeting15:30
*** TrevorV has joined #openstack-meeting15:31
*** bcafarel has quit IRC15:32
*** manpreet has quit IRC15:34
*** bcafarel has joined #openstack-meeting15:36
*** dklyle has joined #openstack-meeting15:48
*** mlavalle has quit IRC15:53
*** mlavalle has joined #openstack-meeting15:55
*** armax has joined #openstack-meeting15:57
*** ociuhandu has joined #openstack-meeting16:02
*** ociuhandu has quit IRC16:09
*** takamatsu has quit IRC16:15
*** jmasud has joined #openstack-meeting16:26
*** baojg has quit IRC16:28
*** baojg has joined #openstack-meeting16:30
*** macz_ has joined #openstack-meeting16:36
*** jmasud has quit IRC16:42
*** takamatsu has joined #openstack-meeting16:46
*** ociuhandu has joined #openstack-meeting16:51
*** jmasud has joined #openstack-meeting16:58
*** hemna has quit IRC16:58
*** TrevorV has quit IRC16:58
*** hemna has joined #openstack-meeting16:59
*** hemna has quit IRC17:08
*** hemna has joined #openstack-meeting17:09
*** ociuhandu_ has joined #openstack-meeting17:15
*** ociuhandu has quit IRC17:17
*** ociuhandu_ has quit IRC17:19
*** jmasud has quit IRC17:26
*** ociuhandu has joined #openstack-meeting17:26
*** ociuhandu has quit IRC17:31
*** e0ne has quit IRC17:36
*** bnemec is now known as beekneemech17:37
*** psachin has quit IRC17:44
*** jmasud has joined #openstack-meeting17:46
*** mlavalle has quit IRC17:58
*** mlavalle has joined #openstack-meeting18:05
*** ralonsoh has quit IRC18:34
*** gyee has joined #openstack-meeting18:54
*** jmasud has quit IRC19:13
*** lifeless has quit IRC19:31
*** lifeless has joined #openstack-meeting19:47
*** slaweq has quit IRC20:03
*** rfolco has quit IRC21:10
*** ociuhandu has joined #openstack-meeting21:19
*** rcernin has joined #openstack-meeting21:19
*** jmasud has joined #openstack-meeting21:32
*** jmasud has quit IRC21:39
*** jmasud has joined #openstack-meeting21:42
*** ociuhandu has quit IRC21:45
*** jmasud has quit IRC21:48
*** jmasud has joined #openstack-meeting21:49
*** jmasud has quit IRC21:54
*** jmasud has joined #openstack-meeting22:15
*** jmasud has quit IRC22:32
*** jmasud has joined #openstack-meeting22:33
*** jmasud has quit IRC22:33
*** jmasud has joined #openstack-meeting22:40
*** rh-jelabarre has quit IRC22:43
*** jmasud has quit IRC22:58
*** jmasud has joined #openstack-meeting22:59
*** jmasud has quit IRC22:59
*** jmasud has joined #openstack-meeting23:09
*** armax has quit IRC23:10
*** armax has joined #openstack-meeting23:13
*** hemna has quit IRC23:19
*** hemna has joined #openstack-meeting23:20
*** corvus has quit IRC23:22
*** rfolco has joined #openstack-meeting23:22
*** jeblair has joined #openstack-meeting23:24
*** jeblair is now known as corvus23:26
*** rfolco has quit IRC23:26
*** tosky has quit IRC23:43

Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!