Thursday, 2016-12-22

*** ducttape_ has joined #openstack-glance00:01
openstackgerritMerged openstack/glance: Add libvirt image metadef for hw_pointer_model  https://review.openstack.org/38613700:17
*** mtanino has quit IRC00:25
*** Zer0Byte__ has quit IRC00:28
*** ducttape_ has quit IRC00:40
*** dillaman has joined #openstack-glance00:42
*** ativelkov_ has joined #openstack-glance00:44
*** ativelkov has quit IRC00:45
*** kragniz has quit IRC00:45
*** jdillaman has quit IRC00:45
*** kragniz has joined #openstack-glance00:46
dharinicrosmaita, sigmavirus, jokke_ , others: So the implementation of CI with the migartion in is: https://review.openstack.org/#/c/369110/ is currently blocked as you all might have already noticed00:49
dharinicThe one failing tempest test was: http://logs.openstack.org/10/369110/57/check/gate-tempest-dsvm-neutron-full-ubuntu-xenial/c118406/console.html#_2016-12-19_20_34_24_37481700:49
dharinicThis was because the test was creating an image with default private visibility.00:49
dharinicstevelle put up a patch that clearly explains the situation. https://review.openstack.org/#/c/412731/00:49
dharinicHowever, the qa team argue that if private visibility is explicitly, specifically mentioned, we are not allowing image member creation and that this is backward incompatible.00:49
dharinicSo unless stevelle’s change goes in, we are blocked on CI. The other way round would be to add atomicity to automatically change the visibility of private images to shared if an attempt to add image members is done.00:49
dharinicPlease refer to the logs in openstack-qa when you get a chance.00:50
stevellebased on discussions in #openstack-qa (timepoint approx Thurs Dec 22 00:50 GMT and before), we might have to at least partially adopt jokke_'s suggestion of auto-migrating private images to shared visibility for $CURRENT_API_VER.  If we do that for all of v2, then the we don't need to merge a change to Tempest tests.00:52
dharinicWe need to decide on what approach we should be taking now with respect to continuing the CI efforts. Either convince the qa team or we have to take the latter path (which is gonna take a while; updating the spec, performing considerable code changes to the CI patch)00:52
stevelleIf we want to cause the microversions to behave differently we would need to auto-migrate private to shared for $current_api_ver00:53
stevelleand that does require a patch merged in Tempest00:55
dharinicstevelle: Can you please brief me as to how and why we decided to do what currently is on the spec as against what jokke_ suggested?00:58
stevelledharinic: there was a mailing list thread and some irc discussion and things settled on not having an implicit behavior00:59
stevelleI can see if I can find you a subject line so you can search the list archives00:59
dharinicsure stevelle. Thanks.00:59
stevelle[openstack-dev] [Glance] [api] Community Images Impasse01:01
stevellethat's the thread01:01
dhariniccool01:01
stevelleactually I think that may be the wrong thread now but not finding what I was thinking of01:15
*** mtanino has joined #openstack-glance01:16
openstackgerritDharini Chandrasekar proposed openstack/glance: Implement and Enable Community Images  https://review.openstack.org/36911001:18
dharinic^ uploaded a new PS for clarity sake, knowing it will fail the tempest test. This has the preferred migration.01:19
stevellegood thing to get out of the way then01:19
dharinicstevelle: Sure. I think there should be one with the API-WG tag?01:22
stevelledoesn't appear to be01:22
*** ducttape_ has joined #openstack-glance01:27
stevelleyeah, I can't find a thread so the discussion I am thinking of must have also been on irc01:35
dharinicstevelle: I very vaguely know about microversioning.. Can you please tell me what the conclusion is reagrding the failing tempest test?01:37
stevelledharinic: you know how we 'bump' the version of the API every so often?01:37
stevellespecifically when a change to the api is made01:37
dharinicYes stevelle01:38
stevelleso, we are bumping the minor version number there.  in theory, (see that link to microversions recommendation from api-wg that mtreinish linked --  http://specs.openstack.org/openstack/api-wg/guidelines/microversion_specification.html -) so the idea is that the client sends a request and can opt to specify the latest minor version of the api it can speak01:40
dharinicahhh gotcha01:41
*** catintheroof has quit IRC01:42
stevelleroughly, what nova does, is maintain multiple implementations of functionality based on that minor version number.  2.1 might accept foo/bar/?baz=splat but 2.2 might not preferring foo/bar/?field=baz&value=splat, so they have to keep code that processes ?baz=splat if the microversion header is 2.101:42
stevellewe must, at the least, continue to accept requests to add a member to a private image if the header is set to the value we used in newton release01:43
dharinicI get the idea.01:43
stevelleand if we do, it has to migrate the visibility to shared01:44
stevelleso that the image will be visible by alt_user in that tempest test01:44
stevelle(just finishing the thought for reference later)01:45
dharinicIts slightly weird that when it is almost the end, there is an issue like this that crops up; there were no serious objections earlier01:45
dharinicHmmm. I understand.01:45
stevelleagreed, and it's because the tempest tests are really testing a boundary condition rather than a common use case01:46
stevelleI might feel lest angsty if they were also testing the common use case :)01:46
dharinicYep. :/01:47
*** ducttape_ has quit IRC03:05
*** btully has quit IRC03:08
*** btully has joined #openstack-glance03:10
*** mtanino has quit IRC03:10
*** btully has quit IRC03:14
stevelledharinic and rosmaita reminder: My end-of-year time off started a few hours ago. I added an agenda item, and tried to make it as complete yet terse as I could. I'll try to make the meeting but I'm never my best at 6am so hoping these notes help the discussion. I tried to be agnostic in my phrasing as well between the options I can envision, but was not perfect.03:17
*** ducttape_ has joined #openstack-glance03:19
*** ducttape_ has quit IRC03:29
*** links has joined #openstack-glance03:42
*** udesale has joined #openstack-glance03:45
*** nicolasbock has quit IRC04:09
*** ducttape_ has joined #openstack-glance04:29
*** pdeore has joined #openstack-glance04:29
*** ducttape_ has quit IRC04:34
*** mtreinish has quit IRC04:44
*** mtreinish has joined #openstack-glance04:45
*** ratailor has joined #openstack-glance05:23
*** ducttape_ has joined #openstack-glance05:31
*** prateek has joined #openstack-glance05:31
*** ducttape_ has quit IRC05:35
*** btully has joined #openstack-glance06:07
*** btully has quit IRC06:15
*** udesale has quit IRC06:17
*** pcaruana has joined #openstack-glance06:17
*** btully has joined #openstack-glance06:40
*** tshefi has joined #openstack-glance06:43
*** openstackgerrit has quit IRC06:48
*** groen692 has joined #openstack-glance06:49
*** e0ne has joined #openstack-glance06:52
*** tesseract has joined #openstack-glance06:58
*** tesseract has quit IRC06:58
*** tesseract has joined #openstack-glance06:59
*** tesseract is now known as Guest4733706:59
*** ezoszed has joined #openstack-glance07:11
*** GB21 has joined #openstack-glance07:25
*** pdeore has quit IRC07:30
*** ducttape_ has joined #openstack-glance07:32
*** pdeore has joined #openstack-glance07:35
*** ducttape_ has quit IRC07:37
*** pdeore has quit IRC07:46
*** pdeore has joined #openstack-glance07:46
*** udesale has joined #openstack-glance08:11
*** GB21 has quit IRC08:18
*** pdeore has quit IRC08:21
*** pdeore has joined #openstack-glance08:21
*** ducttape_ has joined #openstack-glance08:33
*** ducttape_ has quit IRC08:38
*** GB21 has joined #openstack-glance08:38
*** btully has quit IRC08:39
*** btully has joined #openstack-glance08:43
*** GB21 has quit IRC08:44
*** e0ne has quit IRC08:45
*** btully has quit IRC08:47
*** GB21 has joined #openstack-glance08:56
*** zzzeek has quit IRC09:00
*** zzzeek has joined #openstack-glance09:00
*** kuzko has joined #openstack-glance09:03
kuzkohello,09:03
*** GB21 has quit IRC09:04
kuzkoI seem to be triggering an OverflowError for python-glanceclient when downloading an image09:04
kairatHi09:05
kairathow big is the value?09:05
kairator no info in logs?09:05
kairatvalues = image_size09:05
kuzkohttps://zerobin.net/?bfd05129e9029bdf#hOFzRJUn1XdbbHLvDbxG45ufbGN50F0q8Esf+niph1M=09:06
kuzkothat's the only trace of the crash when using --debug09:06
kuzkoimage size = 3GB09:06
kairatoh09:07
kairatAlmost sure there is something with py309:07
kuzkoah... we are a python 3 only staff... is there something I can do about that error?09:08
kairatSo as a workaround you can try to download with py209:08
kuzkook, gotta setup a vm09:09
kairatkuzko, this also might be the problem with requests library09:14
kairatperhaps, sigmavirus may help when he will be available09:14
kuzkokairat  : ok, I'm trying to update my requests libraries and my python env09:16
kairatit seems that all image data was concatenated to response.text with join09:17
kairatso we got overflow error because string becomes too big09:17
kuzkoI'm in a "public cloud user" situation, is it solvable from my side or it need the public cloud provider's eye ?09:19
*** markvoelker has quit IRC09:20
*** GB21 has joined #openstack-glance09:22
kairatso what is your os version?09:26
kairatos = openstack09:26
kuzkoI don't know it's a public cloud named "cloudwatt"09:27
*** mosulica has joined #openstack-glance09:31
kairatkuzko, did you use --debug when download image09:34
kairatit seems that reason is here: https://github.com/openstack/python-keystoneclient/blob/master/keystoneclient/session.py#L21909:35
kairatso when it tries to read big body (like glance image) you have got this error09:37
kairatonce you don't use --debug09:37
kairatit seems to work fine09:37
*** btully has joined #openstack-glance09:42
*** GB21 has quit IRC09:51
*** btully has quit IRC09:53
*** e0ne has joined #openstack-glance09:56
*** hieulq has quit IRC10:07
*** markvoelker has joined #openstack-glance10:21
*** markvoelker has quit IRC10:26
kuzkokairat well after an update of my environment and removing the debug flag, it seems to work. Using the openstackclient works too...10:45
kairatso, it is a bug in keystoneauth1/keystoneclient10:45
kairatwe need to avoid unpacking response if size is big or response is chunked10:46
kairatI guess removing debug did the trick10:46
kairatcause you must still get the error with --debugf10:47
kairatcause you must still get the error with --debug10:47
*** openstackgerrit has joined #openstack-glance10:58
openstackgerritBhagyashri Shewale proposed openstack/python-glanceclient: Fix 'UnicodeEncodeError' for unicode values in url  https://review.openstack.org/32037910:58
jamielennoxif the problem there is that --debug tries to print the body i think that should be fixed in keystoneauth, i don't know if it was ever fixed for keystoneclient.session11:01
jamielennoxif it's still a problem with keystoneauth that should absolutely be a bug11:02
jamielennoxif it's only keystoneclient then i recommend upgrading :)11:02
* jamielennox drops and runs - good night11:03
kairatjamielennox, it seems to  be a bug in keystoneauth https://github.com/openstack/keystoneauth/blob/459f3c0926b1cf62dadafb014b8e7a98e20dc6b9/keystoneauth1/session.py#L35711:03
kairatso not sure if client upgrade will help11:04
kairat* might help11:04
jamielennoxhmm, ok, want to file that as a bug - and maybe fix it? like only log response body if applicatoin/json11:05
jamielennoxor some known list11:05
jamielennoxwe really don't want to try and debug print a full image11:05
kairatyep11:05
jamielennoxi thought this got fixed....11:05
kairatkuzko, could  you create a bug for keystoneauth11:05
*** ducttape_ has joined #openstack-glance11:05
jamielennoxoh, it's kind of fixed for logging a request, but doesn't look like it for logging a response11:06
kuzkokairat : I have no idea how to do so and I don't know if I could reproduce the issue, let me try11:06
jamielennoxreally we need a better logging format in general11:06
jamielennoxbecause https://github.com/openstack/keystoneauth/blob/459f3c0926b1cf62dadafb014b8e7a98e20dc6b9/keystoneauth1/session.py#L337 shouldn't be an indication to dump something into a log11:07
jamielennoxkuzko: well basically if you POST or GET an image or other binary data with --debug we never want to try and print it11:07
kairatkuzko, you can go to https://bugs.launchpad.net/keystoneauth and press create a bug11:07
kairatsorry, report a bug11:08
kairatthen describe your problem11:08
kairatwith the log you posted earlier11:08
kairatafk11:08
kuzkook, I'll do it when I'll be sure I can recreate it, gimme a minute11:08
kairatjr11:08
kairatok11:08
jamielennoxwill leave this with you, but once it's created please ping stevemar and I with it11:08
jamielennoxthanks11:09
*** ducttape_ has quit IRC11:26
*** smatzek has joined #openstack-glance11:30
*** ducttape_ has joined #openstack-glance11:35
*** nicolasbock has joined #openstack-glance11:38
*** pdeore has quit IRC11:43
kuzkook that's very very weird11:43
kuzkosame creds same versions, my colleague has an non iterable without the --debug flag and it seems to work without showing anything with the debug flag11:43
*** btully has joined #openstack-glance11:44
kuzko*non iterable error*11:44
*** ducttape_ has quit IRC11:52
*** btully has quit IRC11:54
stevemarjamielennox: whatchoo talkin bout11:57
stevemarjamielennox: kairat sounds like logged a big response breaks things?11:58
kairatyep11:58
stevemarjamielennox: kairat we had a patch for that: https://review.openstack.org/#/c/359292/ but it needs love11:58
kairatstevemar, we have the same in keystonauth11:59
kairatsee https://github.com/openstack/keystoneauth/blob/459f3c0926b1cf62dadafb014b8e7a98e20dc6b9/keystoneauth1/session.py#L33711:59
kairatresponse.text will try to iterate through all image=)12:00
kairatglance moved to keystoneauth AFAIK so keystoneclient is not so relevant12:00
stevemarkairat: still should fix both parts12:01
stevemarkairat: please check bug https://bugs.launchpad.net/python-keystoneclient/+bug/1616105 -- if you think it's the same add a comment about how it's impacting you12:10
openstackLaunchpad bug 1616105 in python-keystoneclient "Request of large files raises a MemoryError due to logging" [High,In progress] - Assigned to Tobias Diaz (int-0)12:10
stevemarkairat: that is one of the bugs I was hoping to fix over the holidays12:10
kairatOk, will try to post some messages12:10
kairatI see we do not consider chunked-responses in the bug12:11
kairatso we cannot rely on size only12:11
kairatstevemar, ^12:11
*** bkopilov has joined #openstack-glance12:12
*** udesale has quit IRC12:13
*** bkopilov has quit IRC12:18
*** markvoelker has joined #openstack-glance12:22
*** markvoelker has quit IRC12:28
*** lujinluo_ has joined #openstack-glance12:40
*** btully has joined #openstack-glance12:42
*** lujinluo_ has quit IRC12:47
*** ratailor has quit IRC12:49
*** ducttape_ has joined #openstack-glance12:52
*** btully has quit IRC12:54
*** ducttape_ has quit IRC12:57
*** MattMan has quit IRC13:00
*** MattMan has joined #openstack-glance13:00
*** bhagyashris has quit IRC13:09
*** btully has joined #openstack-glance13:12
sigmaviruskairat: you pinged?13:12
kairatsigmavirus, hello! never mind, problem was clarified, I though that the bug is in requests13:16
kairatbut it turns out that it belongs to keystonauth13:16
*** links has quit IRC13:18
abhishek_ksigmavirus: Hi, I have replied to your comments on patch, https://review.openstack.org/#/c/352892/13:22
abhishek_ksigmavirus: please take a look when you get time, thank you13:22
*** chlong has joined #openstack-glance13:23
*** btully has quit IRC13:26
sigmaviruswill do abhishek_k13:31
abhishek_ksigmavirus: thank you13:31
*** btully has joined #openstack-glance13:33
*** btully has quit IRC13:46
*** markvoelker has joined #openstack-glance13:48
*** ducttape_ has joined #openstack-glance13:54
*** ducttape_ has quit IRC13:58
rosmaitaCourtesy meeting reminder on #openstack-meeting-4: ativelkov, cpallares, flaper87, flwang1, hemanthm, jokke_, kragniz, lakshmiS, mclaren, mfedosin, nikhil_k, Nikolay_St, Olena, pennerc, rosmaita, sigmavirus24, sabari, TravT, ajayaa, GB21, bpoulos, harshs, abhishek, bunting, dshakhray, wxy, dhellmann, kairat, aavraham, alex_bash13:58
*** pdeore has joined #openstack-glance14:00
*** mosulica has quit IRC14:03
*** ducttape_ has joined #openstack-glance14:19
*** prateek has quit IRC14:21
*** smatzek has quit IRC14:30
*** ducttape_ has quit IRC14:32
*** btully has joined #openstack-glance14:35
*** bkopilov has joined #openstack-glance14:47
*** ducttape_ has joined #openstack-glance14:56
*** ezoszed has quit IRC15:02
*** mtanino has joined #openstack-glance15:03
dharinicokay, so regarding the PATCH bug, i tried to incorporate what was missing with respect to the bug reported. As per rfc6902. Was not clear as to what kairat pointed out in his review in that case. Would like some inputs on this.15:03
dharinicMeanwhile, I will also look more into this.15:03
dharinicreference: https://bugs.launchpad.net/glance/+bug/1521607 , https://review.openstack.org/#/c/401391/15:04
openstackLaunchpad bug 1521607 in Glance "v2 - replacing array elements with PATCH results in 400 error" [Medium,In progress] - Assigned to Dharini Chandrasekar (dharini-chandrasekar)15:04
dharinicstevelle, rosmaita, sigmavirus^15:08
dharinicThanks rosmaita15:09
*** smatzek has joined #openstack-glance15:13
*** e0ne has quit IRC15:14
*** ducttape_ has quit IRC15:19
*** ducttape_ has joined #openstack-glance15:26
rosmaitadharinic: hope my comments on the review helped15:28
*** mtan_____ has joined #openstack-glance15:34
*** pdeore has quit IRC15:35
*** mtanino has quit IRC15:35
*** ducttape_ has quit IRC15:41
sigmavirusflaper87: you around?15:53
stevellerosmaita: I have a new review staged for tempest, ready to refer to the list discussion in it15:59
rosmaitastevelle: ok, thanks, i am going to be closer to 16:30 with my draft, sorry16:00
*** Guest47337 has quit IRC16:03
*** ducttape_ has joined #openstack-glance16:15
flaper87sigmavirus: I am16:16
rosmaitaflaper87: happy holidays!16:16
flaper87rosmaita: happy holidays for you too :D16:17
rosmaitaflaper87: what i got for christmas was tempest tests breaking because of community images16:17
rosmaitalook for a ML conversation16:17
rosmaitastevelle has a patch up to change the tempest test, but some people on the patch are worried about backward compatibility16:18
flaper87rosmaita: ooh crap :(16:18
flaper87that doesn't sound very good :(16:18
* flaper87 sad panda16:18
*** tshefi has quit IRC16:29
*** ducttape_ has quit IRC16:34
sigmavirusflaper87: happy holidays! May I PM you?16:34
flaper87sigmavirus: please do16:34
*** smatzek has quit IRC16:44
dharinicrosmaita: It sure did. Will investigate if we can use a library directly. If not, will leave my patch to be reviewed as it is.16:50
dharinicthanks :)16:50
*** groen692 has quit IRC17:03
*** ducttape_ has joined #openstack-glance17:05
*** btully has quit IRC17:05
*** bkopilov has quit IRC17:06
stevelleis there a preferred way to deal with long urls in commit messages?17:07
rosmaitastevelle: not sure, sigmavirus ^^17:08
*** bkopilov has joined #openstack-glance17:12
sigmavirusstevelle: just put them there17:13
sigmavirusURLs (like in code) are exemptions17:13
stevellemaking sure.17:13
sigmavirus+117:14
rosmaitasigmavirus: ty17:14
sigmavirusHTH17:14
stevellethe next thing is will I need to use the addCleanup hook for a resource created in the course of test execution, but not during setup. leaning toward no17:14
stevellethis is in tempest, so use that filter for correctness17:15
*** ducttape_ has quit IRC17:18
*** flaper87 has quit IRC17:22
*** btully has joined #openstack-glance17:31
*** flaper87 has joined #openstack-glance17:35
*** btully has quit IRC17:42
rosmaitastevelle: alex_bash: sigmavirus: i have to grab lunch, back in 20 min & will try to be inspired by food to have a nice, snappy ending17:56
*** nicolasbock has quit IRC18:01
*** nicolasbock has joined #openstack-glance18:06
openstackgerritMaxim Nestratov proposed openstack/glance: Add ploop to supported disk_formats  https://review.openstack.org/34163318:08
openstackgerritMaxim Nestratov proposed openstack/glance: Add ploop to supported disk_formats  https://review.openstack.org/34163318:14
*** ducttape_ has joined #openstack-glance18:18
*** ducttape_ has quit IRC18:23
openstackgerritMerged openstack/glance: Fix some typos in api-ref  https://review.openstack.org/41365218:45
openstackgerritOpenStack Proposal Bot proposed openstack/glance: Updated from global requirements  https://review.openstack.org/41426618:48
dharinicbad tempest behaviours in glance_store: http://logs.openstack.org/66/120866/20/check/gate-tempest-dsvm-neutron-src-glance_store-ubuntu-xenial/6216c32/console.html18:50
dharinicDo we report a bug or something?18:51
rosmaitadharinic: you might want to put a shout in #openstack-infra and ask someone to take a look19:05
rosmaita... which i now see you already did!19:06
*** ducttape_ has joined #openstack-glance19:06
stevelle#kudo dharinic19:07
dharinicThanks stevelle and rosmaita19:10
rosmaitastevelle: alex_bash: dharinic: sigmavirus: jokke_: last call for comments on https://etherpad.openstack.org/p/glance-ocata-community-images-tempest-imbroglio19:15
rosmaitamy plan is to send it out in 1/2 hour19:15
dharinicLooks good to me rosmaita19:23
*** ducttape_ has quit IRC19:23
*** rcernin has joined #openstack-glance19:38
openstackgerritDharini Chandrasekar proposed openstack/glance_store: swift store: Add upload recovery  https://review.openstack.org/12086619:46
rosmaitastevelle: http://lists.openstack.org/pipermail/openstack-dev/2016-December/109370.html19:57
stevellethanks, cross-linked19:58
stevellehope that copy-paste worked, I didn't even open it yet :P19:59
*** e0ne has joined #openstack-glance20:05
*** ducttape_ has joined #openstack-glance20:24
*** e0ne has quit IRC20:27
*** ducttape_ has quit IRC20:29
*** e0ne has joined #openstack-glance20:30
*** e0ne has quit IRC20:41
*** mtan_____ has quit IRC21:09
*** btully has joined #openstack-glance21:16
*** btully has quit IRC21:21
*** ducttape_ has joined #openstack-glance21:25
*** ducttape_ has quit IRC21:30
*** mvk has quit IRC21:42
*** ChanServ changes topic to "OpenStack Glance"21:57
*** ChanServ changes topic to "OpenStack Glance | priorities: http://lists.openstack.org/pipermail/openstack-dev/2016-December/109375.html"21:57
*** ChanServ changes topic to "OpenStack Glance | priorities: http://lists.openstack.org/pipermail/openstack-dev/2016-December/109375.html | logs at http://eavesdrop.openstack.org/irclogs/%23openstack-glance/"21:58
*** seanhandley has quit IRC22:07
*** seanhandley has joined #openstack-glance22:24
*** btully has joined #openstack-glance22:26
*** ducttape_ has joined #openstack-glance22:26
*** ducttape_ has quit IRC22:31
*** chlong has quit IRC22:34
*** Zer0Byte__ has joined #openstack-glance22:34
*** btully has quit IRC22:36
*** chlong has joined #openstack-glance22:41
*** Zer0Byte__ has quit IRC22:46
*** ducttape_ has joined #openstack-glance22:54
*** Zer0Byte__ has joined #openstack-glance23:13
*** ducttape_ has quit IRC23:25
*** ducttape_ has joined #openstack-glance23:59

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!