Thursday, 2020-12-10

*** tosky has quit IRC00:02
*** baojg has quit IRC00:18
*** baojg has joined #openstack-cinder00:19
*** enriquetaso has quit IRC00:31
*** brinzhang has joined #openstack-cinder00:38
*** benj_- has joined #openstack-cinder00:53
*** benj_ has quit IRC00:54
*** benj_- is now known as benj_00:54
lsekigeguileo: indeed it requires implementing update_migrated_volume, too01:07
lsekinow the QoS policy group is associated to the correct file name01:08
*** thgcorrea has quit IRC01:29
*** user_19173783170 has joined #openstack-cinder01:44
*** user_19173783170 has left #openstack-cinder02:16
*** user_19173783170 has joined #openstack-cinder02:16
*** priteau has quit IRC03:03
*** openstackgerrit has quit IRC03:22
*** psachin has joined #openstack-cinder03:44
*** hamalq_ has quit IRC03:45
*** amar7ibm_ has joined #openstack-cinder03:53
*** dave-mccowan has quit IRC03:54
*** rosmaita has left #openstack-cinder03:57
*** openstackgerrit has joined #openstack-cinder04:28
openstackgerritMasahiro Okuri proposed openstack/cinder master: created s3 cinder backup driver  https://review.opendev.org/c/openstack/cinder/+/74656104:28
*** sapd1 has joined #openstack-cinder04:55
*** zzzeek has quit IRC05:06
*** zzzeek has joined #openstack-cinder05:08
*** dasp has quit IRC05:18
*** sharathkacham has joined #openstack-cinder05:26
*** user_19173783170 has quit IRC05:26
*** udesale has joined #openstack-cinder05:30
*** dasp has joined #openstack-cinder05:34
*** user_19173783170 has joined #openstack-cinder05:38
*** user_19173783170 has quit IRC05:42
*** rpittau|afk has quit IRC06:11
*** mnaser has quit IRC06:11
*** mnaser has joined #openstack-cinder06:11
*** rpittau|afk has joined #openstack-cinder06:12
*** sapd1 has quit IRC06:15
*** zzzeek has quit IRC06:17
*** haixin has joined #openstack-cinder06:17
*** zzzeek has joined #openstack-cinder06:18
*** pcaruana has joined #openstack-cinder06:19
*** zenkuro has joined #openstack-cinder06:31
*** GirishChilukuri has joined #openstack-cinder06:40
*** GirishChilukuri has quit IRC06:45
*** lpetrut has joined #openstack-cinder07:00
*** rcernin has quit IRC07:10
*** rcernin has joined #openstack-cinder07:29
*** zenkuro has quit IRC07:58
*** zenkuro has joined #openstack-cinder07:58
*** hamalq has joined #openstack-cinder08:06
*** rcernin has quit IRC08:07
*** rpittau|afk is now known as rpittau08:17
*** e0ne has joined #openstack-cinder08:23
*** zzzeek has quit IRC08:36
*** zzzeek has joined #openstack-cinder08:39
*** tosky has joined #openstack-cinder08:47
*** zenkuro has quit IRC08:49
*** zenkuro has joined #openstack-cinder08:49
*** spatel has joined #openstack-cinder08:52
*** spatel has quit IRC08:56
*** rcernin has joined #openstack-cinder08:56
*** dsariel has joined #openstack-cinder09:01
*** haixin has quit IRC09:12
*** rcernin has quit IRC09:13
*** martinkennelly has joined #openstack-cinder09:16
*** zzzeek has quit IRC09:32
*** zzzeek has joined #openstack-cinder09:33
*** udesale_ has joined #openstack-cinder10:05
*** hamalq has quit IRC10:07
*** udesale has quit IRC10:08
*** zzzeek has quit IRC10:13
*** zzzeek has joined #openstack-cinder10:14
openstackgerritLucian Petrut proposed openstack/os-brick master: Update requirements  https://review.opendev.org/c/openstack/os-brick/+/76644210:26
*** zzzeek has quit IRC10:39
*** ociuhandu has joined #openstack-cinder10:40
*** zzzeek has joined #openstack-cinder10:41
*** ociuhandu has quit IRC10:54
*** rcernin has joined #openstack-cinder10:57
*** sumanshaw1 has joined #openstack-cinder11:04
*** hamalq has joined #openstack-cinder11:09
*** hamalq has quit IRC11:14
sumanshaw1geguileo: Hello11:17
*** ociuhandu has joined #openstack-cinder12:02
*** ociuhandu has quit IRC12:03
*** ociuhandu has joined #openstack-cinder12:03
*** sumanshaw1 has quit IRC12:03
*** GirishChilukuri has joined #openstack-cinder12:13
*** rcernin has quit IRC12:13
*** zzzeek has quit IRC12:14
*** zzzeek has joined #openstack-cinder12:17
*** ociuhandu has quit IRC12:23
*** lxkong has quit IRC12:26
*** hamalq has joined #openstack-cinder12:28
*** ociuhandu has joined #openstack-cinder12:31
*** hamalq has quit IRC12:33
*** TusharTgite has joined #openstack-cinder12:33
*** zzzeek has quit IRC12:40
*** GirishChilukuri has quit IRC12:40
*** zzzeek has joined #openstack-cinder12:41
*** baojg has quit IRC12:46
*** baojg has joined #openstack-cinder12:47
*** ociuhandu has quit IRC12:55
*** dave-mccowan has joined #openstack-cinder12:57
*** benfelin has joined #openstack-cinder12:59
*** rosmaita has joined #openstack-cinder13:04
*** ociuhandu has joined #openstack-cinder13:05
*** priteau has joined #openstack-cinder13:10
*** ociuhandu has quit IRC13:15
openstackgerritLucian Petrut proposed openstack/os-brick master: Update requirements  https://review.opendev.org/c/openstack/os-brick/+/76644213:20
*** hemna has quit IRC13:28
*** _pewp_ has quit IRC13:29
*** _pewp_ has joined #openstack-cinder13:30
*** hemna has joined #openstack-cinder13:31
*** zenkuro has quit IRC13:35
*** zenkuro has joined #openstack-cinder13:36
*** _erlon_ has joined #openstack-cinder13:50
*** zigo has joined #openstack-cinder13:51
*** ociuhandu has joined #openstack-cinder13:52
rosmaitalpetrut: thank you!!!13:54
lpetrutrosmaita: no problem, I hope I got it right this time :)13:55
lpetrutrosmaita: we'll probably have to improve the tooling around lower constraints at some point13:56
rosmaitalpetrut: or eliminate them altogether -- i'm not sure what kind of value they are providing13:56
rosmaita(other than that they forced us to update all our requirements, some of which were extremely out of date)13:57
lpetrutI guess there's value in ensuring that both upper as well as the lower constraints work, otherwise the services might fail at runtime due to invalid dependencies13:58
lpetrutI'm not sure though if indirect dependencies should be set in the lower constraints file13:58
lpetrutthe docs aren't very clear on that either13:59
*** sapd1 has joined #openstack-cinder13:59
*** sapd1 has quit IRC14:00
rosmaitayes, there was some discussion on that back when the lower-constraints job was introduced, but i don't remember the reasoning any more14:00
*** sapd1 has joined #openstack-cinder14:01
*** enriquetaso has joined #openstack-cinder14:03
*** ociuhandu has quit IRC14:06
*** ociuhandu has joined #openstack-cinder14:09
*** ociuhandu has quit IRC14:13
*** ociuhandu has joined #openstack-cinder14:15
*** ociuhandu has quit IRC14:19
*** psachin has quit IRC14:28
*** ociuhandu has joined #openstack-cinder14:28
*** hamalq has joined #openstack-cinder14:29
*** hamalq_ has joined #openstack-cinder14:32
*** hamalq has quit IRC14:34
*** hamalq_ has quit IRC14:36
*** tobberydberg has quit IRC14:38
*** tobberydberg has joined #openstack-cinder14:39
smcginnisYes, indirect dependencies should be there. The idea was to capture and constrain the whole set of dependencies that get installed to make sure we could convey an accurate description of the range of versions expected to work when packaged together.14:47
rosmaitasmcginnis: i guess that means we should occasionally review the l-c file as some indirect dependencies are dropped & added?14:49
rosmaitais pip freeze on a testenv the best way to figure that out?14:49
*** ociuhandu has quit IRC14:57
smcginnisrosmaita: Yeah, I guess so.14:58
smcginnisRecognizing when to bump up an l-c is a little trickier. You need to pay attention to when something needs to be raised for a given feature.14:58
smcginnisBut indirect dependencies and whether to clean out things is a fairly easy check I think.14:59
rosmaitawell, i bumped them all the way up to our requirements minimums14:59
smcginnisYou can do "tox -re lower-constraints --notest; .tox/lower-constraints/bin/pip freeze > logfile" and compare the contents of logfile with lower-constraints.txt.14:59
rosmaitasome of them were 2-3 major versions behind15:00
smcginnisAnd that's after one or two patches to bring them at least sort of up to date. I'm sure the stable branch versions for something like train are probably pretty bad.15:01
rosmaitasmcginnis: yeah, that's basically what i did15:01
rosmaitai can only imagine!15:01
rosmaitai haven't looked to see the state of our victoria gate15:01
smcginnisFrom the amount of emails I've been getting from nightly stable jobs, I think every stable branch is broken right now.15:02
*** andrebeltrami has joined #openstack-cinder15:02
*** ociuhandu has joined #openstack-cinder15:05
*** e0ne has quit IRC15:07
*** e0ne has joined #openstack-cinder15:08
*** ociuhandu has quit IRC15:10
*** ociuhandu has joined #openstack-cinder15:11
e0negeguileo: hi Gorka. are you around?15:11
geguileoe0ne: yup, on a meeting, so slow response15:12
e0neI'm curious how to reproduce this https://bugs.launchpad.net/cinder/+bug/1801958 issue15:13
openstackLaunchpad bug 1801958 in Cinder "Service down and errors in log during image related operations" [Undecided,Fix released] - Assigned to Jay Bryant (jsbryant)15:13
e0negeguileo: with your fix 8GB image download process takes 10 minutes instead of 2-315:14
*** ociuhandu has quit IRC15:15
geguileoe0ne: if we can fix the eventlet bug with logging we could get it down back to 2-315:15
geguileoe0ne: but we cannot log in a native thread from the pool  :-(15:16
geguileoe0ne: I can't tell in which specific cases the I/O has the problem15:16
geguileoe0ne: I've seen it happen, and using native threads fix it15:17
e0negeguileo: it's true, but your change doesn't do anything with logs. it spawns threads during file read/write15:17
e0negeguileo: I can't reproduce the original issue now15:18
geguileoe0ne: I meant that if that bug didn't exist I wouldn't need to call tpool.execute for each _verify_image call15:18
geguileoe0ne: I could just run the whol verify_glance_image_signature method in a native pool15:18
geguileos/pool/thread15:18
geguileobut if we call the whole verify_glance_image_signature in a native thread, it will eventually block and hang forever15:19
geguileoe0ne: if we could fix the logging issue we could improve Cinder's performance considerably  :-(15:20
*** ociuhandu has joined #openstack-cinder15:20
e0negeguileo: we're talking about different things15:21
geguileoe0ne: how so?15:22
e0negeguileo: I agree with an issue with verify_glance_image_signature (I didn't verify it, but that sounds reasonable)15:22
e0negeguileo: https://review.opendev.org/c/openstack/cinder/+/615934/2/cinder/image/image_utils.py - line #37715:23
e0netpool.Proxy(image_file))15:23
e0neit will spawn a new thread each time we write image chunk to file15:23
e0neand there is no logging in this call15:23
geguileoe0ne: same thing, if we could use logs on a native thread we would not need to proxy the object and we could just proxy the whole fetch15:24
potsHi all.  The Dell EMC PowerVault ME driver is passing CI and ready for reviews: https://review.opendev.org/c/openstack/cinder/+/758684.  Should be an easy one as this is it just inherits all its classes from an existing driver.  Thanks!15:25
e0negeguileo: but why do we need to spawn process per each write call here https://github.com/openstack/cinder/blob/master/cinder/image/glance.py#L380-L381?15:25
geguileoe0ne: that needs to be called on a native thread to prevent it from blocking other greethreads15:26
e0newhy?15:26
e0neeventlet should patch such I/O operations15:26
geguileoe0ne: because sometimes I/O calls don't switch15:26
openstackgerritWalt proposed openstack/cinder master: Add ceph iscsi volume driver  https://review.opendev.org/c/openstack/cinder/+/66282915:27
geguileoe0ne: it should, but it some cases it doesn't seem to behave that way15:27
geguileoe0ne: which is the problem that patch tries to solve15:27
e0negeguileo: do you know how to reproduce such issue?15:27
geguileothose weird cases15:27
geguileoe0ne: unfortunately no15:27
e0ne:(15:28
geguileoe0ne: I didn't like it, but that's why I had to write "in some cases block greenthreads" in the commit message and couldn't say which cases those were15:28
geguileoe0ne: iirc I think swift had found a way to work around the logging issue15:29
e0negeguileo: it's very strange that evenlet blocks on such I/O tasks15:29
e0neit's not related to logging15:30
geguileoe0ne: but if we can fix the problem of logging into a native thread when using greenthreads our performance issue would be resolved15:31
hemnarosmaita thanks for fixing the lower-constraints mess15:31
e0negeguileo: actually, there is an issue with performance which coud be fixed now, I'll propose a patch15:31
rosmaitahemna: thanks for kicking off the process!15:31
geguileoe0ne: nice!!15:33
*** TusharTgite has quit IRC15:43
*** sapd1 has quit IRC15:52
openstackgerritIvan Kolodyazhny proposed openstack/cinder master: Image download: do not write to file in a native thread  https://review.opendev.org/c/openstack/cinder/+/76650115:57
e0negeguileo: here is it ^^^15:57
*** lpetrut has quit IRC15:57
geguileoe0ne: I think we cannot do that...15:59
e0newhy?15:59
geguileoe0ne: because for RBD we will get a handle, which is handled by C code16:00
geguileoand it will block16:00
geguileoiirc16:00
e0nerbd?16:00
geguileoos-brick's rbd connector returns a file handle, right?16:00
geguileowait, maybe I'm mixing things16:01
e0neno16:01
*** sapd1 has joined #openstack-cinder16:01
geguileoI'll have a better look at the patch16:01
e0neit's two lines above16:01
e0nethanks16:01
* geguileo is still in the never-ending meeting16:01
*** sapd1 has quit IRC16:07
*** pcaruana has quit IRC16:10
toskygeguileo: sorry :/16:13
*** e0ne has quit IRC16:28
openstackgerritHelen Walsh proposed openstack/cinder master: mypy - cinder driver.py  https://review.opendev.org/c/openstack/cinder/+/76581016:29
*** hamalq has joined #openstack-cinder16:30
*** ociuhandu has quit IRC16:31
*** ociuhandu has joined #openstack-cinder16:36
*** ociuhandu has quit IRC16:48
*** ociuhandu has joined #openstack-cinder16:53
*** ociuhandu_ has joined #openstack-cinder16:59
openstackgerritBrian Rosmaita proposed openstack/cinder master: Add cinder permissions matrix  https://review.opendev.org/c/openstack/cinder/+/76330617:00
*** gksk has joined #openstack-cinder17:03
*** ociuhandu has quit IRC17:03
gkskTo the core reviewers please review the below feature related to revert to snapshot17:04
gkskhttps://review.opendev.org/c/openstack/cinder/+/75183717:04
*** sharathkacham has quit IRC17:05
geguileotosky: lol, not your fault, it's the job  XD17:05
*** manoj_kumar_kata has joined #openstack-cinder17:05
*** gksk has quit IRC17:09
*** dave-mccowan has quit IRC17:33
*** hamalq has quit IRC17:34
*** dave-mccowan has joined #openstack-cinder17:34
*** udesale_ has quit IRC17:34
*** dave-mccowan has quit IRC17:39
*** ociuhandu_ has quit IRC17:42
*** ociuhandu has joined #openstack-cinder17:54
*** ociuhandu has quit IRC17:58
*** hamalq has joined #openstack-cinder18:03
*** rpittau is now known as rpittau|afk18:04
*** hamalq_ has joined #openstack-cinder18:07
*** hamalq has quit IRC18:10
*** CeeMac has joined #openstack-cinder18:34
*** _pewp_ has quit IRC19:20
*** _pewp_ has joined #openstack-cinder19:21
*** _pewp_ has quit IRC19:31
*** _pewp_ has joined #openstack-cinder19:31
*** benfelin has quit IRC19:43
*** martinkennelly has quit IRC20:06
*** benfelin has joined #openstack-cinder20:07
*** martinkennelly has joined #openstack-cinder20:10
openstackgerritRafael Weingartner proposed openstack/cinder master: Introduces MV to add volume type ID in volumes details  https://review.opendev.org/c/openstack/cinder/+/66688620:10
*** martinkennelly has quit IRC20:19
*** amar7ibm_ has quit IRC20:20
*** rcernin has joined #openstack-cinder20:21
*** martinkennelly has joined #openstack-cinder20:27
openstackgerritAlan Bishop proposed openstack/python-cinderclient master: Update requirements and lower-constraints  https://review.opendev.org/c/openstack/python-cinderclient/+/76654420:43
eharneyis "ERROR Failed to update project None in 5s" a known issue on our CI jobs?20:53
eharneyhas happened a few times on  https://review.opendev.org/c/openstack/cinder/+/73362120:53
openstackgerritAlan Bishop proposed openstack/python-cinderclient master: Update requirements and lower-constraints  https://review.opendev.org/c/openstack/python-cinderclient/+/76654420:53
clarkbeharney: yes, the tldr is that at least one of the zuul executors failed to clone nova properly (due to timeouts that were too aggressive) which led it to being unable to configure jobs with nova in them20:53
clarkbeharney: we've since fixed that20:53
eharneyclarkb: ahh ok, thanks20:53
openstackgerritEric Harney proposed openstack/cinder master: mypy: annotate image/glance.py  https://review.opendev.org/c/openstack/cinder/+/76606421:00
*** lxkong has joined #openstack-cinder21:02
openstackgerritTom Barron proposed openstack/devstack-plugin-ceph master: Tweak ganesha repo cleanup  https://review.opendev.org/c/openstack/devstack-plugin-ceph/+/76655021:13
*** rcernin has quit IRC21:31
*** rcernin has joined #openstack-cinder21:34
*** eharney has quit IRC21:38
*** ociuhandu has joined #openstack-cinder21:40
*** rcernin has quit IRC21:51
*** rcernin has joined #openstack-cinder21:53
*** eharney has joined #openstack-cinder22:02
*** andrebeltrami has quit IRC22:11
*** rosmaita has left #openstack-cinder22:13
*** enriquetaso has quit IRC22:39
*** martinkennelly has quit IRC22:48
*** ociuhandu has quit IRC22:59
openstackgerritAlan Bishop proposed openstack/python-cinderclient master: Update requirements and lower-constraints  https://review.opendev.org/c/openstack/python-cinderclient/+/76654423:00
*** smcginnis has quit IRC23:16
*** baojg has quit IRC23:22
*** baojg has joined #openstack-cinder23:25
*** smcginnis has joined #openstack-cinder23:42
*** ChanServ sets mode: +o smcginnis23:42
*** smcginnis has quit IRC23:49
openstackgerritSam Morrison proposed openstack/cinder master: Make availability zone type exception easier to understand.  https://review.opendev.org/c/openstack/cinder/+/76487523:51
*** baojg has quit IRC23:52
*** baojg has joined #openstack-cinder23:53
*** tosky has quit IRC23:53

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