Friday, 2017-07-28

openstackgerritOpenStack Proposal Bot proposed openstack/manila-ui master: Updated from global requirements
openstackgerritMerged openstack/python-manilaclient master: Enable IPv6 in manila(client)
openstackgerritMerged openstack/manila master: Updated from global requirements
*** jiaopengju has joined #openstack-manila00:32
*** jiaohaolin1 has quit IRC00:42
*** catintheroof has quit IRC00:57
*** harlowja has quit IRC01:10
*** kaisers__ has joined #openstack-manila01:17
*** catintheroof has joined #openstack-manila01:18
*** kaisers_ has quit IRC01:20
*** catintheroof has quit IRC01:25
openstackgerritzhongjun proposed openstack/manila master: Fix tempests tests cannot run in test_rules
*** jiaopengju has quit IRC01:30
*** jiaopengju has joined #openstack-manila01:31
*** gouthamr has quit IRC01:33
openstackgerritzhongjun proposed openstack/manila master: Enable IPv6 in manila(documentation)
*** jmlowe has joined #openstack-manila02:04
*** gouthamr has joined #openstack-manila03:22
*** gouthamr has quit IRC03:46
*** gouthamr has joined #openstack-manila03:58
*** gouthamr has quit IRC03:58
*** gouthamr has joined #openstack-manila03:59
openstackgerritOpenStack Proposal Bot proposed openstack/manila master: Updated from global requirements
*** ganso has quit IRC04:24
*** gouthamr has quit IRC04:30
*** dsariel has joined #openstack-manila04:30
*** harlowja has joined #openstack-manila04:32
*** arnewiebalck_ has joined #openstack-manila04:41
*** kaisers__ has quit IRC05:19
*** dsariel has quit IRC05:26
*** harlowja has quit IRC05:28
*** kaisers_ has joined #openstack-manila05:32
*** kaisers_ has quit IRC05:37
*** harlowja has joined #openstack-manila05:41
*** arnewiebalck_ has quit IRC05:47
*** kaisers_ has joined #openstack-manila05:50
*** sticker_ has quit IRC06:23
*** kaisers_ has quit IRC06:43
*** harlowja has quit IRC06:48
*** kaisers has quit IRC07:20
*** kaisers has joined #openstack-manila07:29
*** kaisers has quit IRC07:37
*** kaisers has joined #openstack-manila07:39
*** kaisers1 has joined #openstack-manila07:42
*** kaisers has quit IRC07:50
*** zhonghua2 has quit IRC08:07
*** zhonghua2 has joined #openstack-manila08:07
*** openstackgerrit has quit IRC08:18
*** tuanluong has joined #openstack-manila08:45
*** kaisers has joined #openstack-manila08:47
*** kaisers has quit IRC09:24
*** gcb has quit IRC09:53
*** ganso has joined #openstack-manila09:54
*** jiaopengju has quit IRC10:06
*** openstackgerrit has joined #openstack-manila10:10
openstackgerritTatyana proposed openstack/manila master: [do not merge] manila CHEPH shares tests
*** kaisers has joined #openstack-manila10:21
*** dsariel has joined #openstack-manila10:24
*** tuanluong has quit IRC10:31
*** kaisers has quit IRC10:32
openstackgerritOpenStack Proposal Bot proposed openstack/manila-ui master: Imported Translations from Zanata
*** kaisers has joined #openstack-manila11:34
*** catintheroof has joined #openstack-manila11:48
*** catintheroof has quit IRC12:14
openstackgerritTom Barron proposed openstack/manila-ui master: Fix gate manilclient import error
*** catintheroof has joined #openstack-manila12:50
*** kaisers has quit IRC12:51
*** Yogi1 has joined #openstack-manila12:53
gansotbarron: nice!12:54
tbarronganso: think that will work12:55
gansotbarron: according to the error in the logs, I don't think so12:56
gansotbarron: I saw errors failing to import availability_zone12:56
gansotbarron: so I guess the manilaclient version in the requirements needs to be update12:56
gansotbarron: I had this problem back when we merged mountable snapshots12:57
tbarronganso: it works locally.  from 'manilaclient.v2 import availability_zones' fails because 'import manilaclient.v2' fails12:58
tbarronganso: hmm, so let me look at requirements again12:58
gansotbarron: oh, and manilaclient.v2 fails because of ipaddress?12:58
gansotbarron: I remember that we bumped upper constraints to update manila-ui12:59
gansotbarron: I am not sure if this is necessary in this case12:59
gansotbarron: haven't looked into it much12:59
tbarronyes, manilaclient.v2 fails b/c of ipaddress13:00
tbarronimporting ipaddress fixes the problem locally13:00
gansotbarron: oh, cool, so your fix will most likely work13:00
tbarroni mean: requiring ipaddress13:00
tbarronbut let me look at the manilaclient versiono13:01
tbarronversion too13:01
tbarronthat might be a better way13:01
*** raissa_ is now known as raissa13:03
tbarronganso: heh, manila-ui is requiring python-manilaclient >= 1.12.0 and we just released 1.17.013:03
raissatbarron: the voting gates seem happy now (I think errors in nv gates are unrelated)
*** xyang1 has joined #openstack-manila13:06
tbarronraissa: nice!13:08
tbarronraissa: cephfs-nfs is broken b/c of new luminous minor version; rraja is looking at it13:08
raissaahh all right :)13:08
tbarronraissa: and the other non-votings like hdfs and glusterfs-native have been failing "forever"13:09
tbarronthey are basically unsupported atm13:09
tbarronand cephfs-native job just timed out, slow devstack load, an intermittent issue for all jobs13:10
raissaandreaf: ^ re patch 33459613:10
tbarronraissa: so I agree that the nonvoting job failures have nothing to do with your patch13:10
andreafraissa: nice work13:11
*** eharney has joined #openstack-manila13:14
openstackgerritRaissa Sarmento proposed openstack/manila master: Use tempest-plugin service client registration
gansotbarron: we need to update the upper constraints then I suppose13:20
tbarronganso: looks like it has 1.15.0 for python-manilaclient13:23
*** dustins has joined #openstack-manila13:24
tbarronganso: we also have explicit requirement line in manila-ui/requirements, I just tried bumping it locally and that alone doesn't fix the issue13:24
tbarronganso: but offhand I don't see why both manila-ui/requirements and upper-constraints shouldn't be 1.17.0 now13:25
openstackgerritTom Barron proposed openstack/manila-ui master: DNM - requirements bump expt
bswartzguys I created the pike-rc1 milestone yesterday in LP13:33
bswartzplease target bugs to that13:33
tbarronraissa: please add yourself as co-author on 33459613:34
bswartztbarron: u-c should be 1.17 but requirements doesn't need to move unless something is broken13:34
tbarronbswartz: ack, something is broken but fixable by requiring ipaddress module in manila-ui w/o bumping python-manilclient requirement13:35
tbarronbswartz: and bumping the requirement doesn't itself fix the issue13:35
gansotbarron: jenkins failed13:36
bswartzcan someone explains to me how the import of "ipaddress" could possibly fail? isn't that part of the standard library?13:36
tbarronbswartz: good question13:36
tbarroni think it is starting in python 2.6 or 2.713:36
tbarronand of course in python313:36
tbarronganso: i didn't specify a version on ipaddr13:38
bswartzI don't understand how parts of the standard library could have versions -- it's just the python version right?13:38
tbarronrequirement, looks like I need to make it match, i'll set it to >=1.0.713:38
bswartzam I wrong about ipaddress being in the standard library?13:38
tbarronbswartz: maybe not in python2?13:39
openstackgerritRaissa Sarmento proposed openstack/manila master: Use tempest-plugin service client registration
tbarronbswartz: we aren't having to include it in requirements in manila or python-manilaclient13:39
tbarronso either we have a transitive requirement there that makes it work in python2 or something13:40
tbarronfunny is going on in manila-ui13:40
bswartzthe sooner we kill the py2.7 the better13:40
*** cknight has joined #openstack-manila13:41
openstackgerritTom Barron proposed openstack/manila-ui master: Fix gate manilclient import error
*** kaisers has joined #openstack-manila13:48
*** gouthamr has joined #openstack-manila13:54
*** tosky has joined #openstack-manila13:59
*** tellesnobrega has joined #openstack-manila14:00
toskytbarron: hi! I've noticed and the fix, we have a similar issue in sahara-dashboard14:00
openstackLaunchpad bug 1707194 in manila-ui "test: gate: Failure: ImportError (cannot import name availability_zones)" [Critical,In progress] - Assigned to Tom Barron (tpb)14:00
*** kaisers has quit IRC14:00
toskytbarron: but shouldn't the import be added to manilaclient, instead of each possible users of manilaclient (in this case, manila-ui, but also any other?)14:00
tbarrontosky: probably right, please -1 that patch14:01
* tbarron goes off to BJ14:01
tbarrontosky: and feel free to propose the right fix in the mean time14:01
toskyack :)14:01
openstackgerritLuigi Toscano proposed openstack/python-manilaclient master: Fix manilclient.v2 import error
openstackgerritLuigi Toscano proposed openstack/python-manilaclient master: Fix manilclient.v2 import error
*** AlexeyAbashkin has joined #openstack-manila14:18
openstackgerritLuigi Toscano proposed openstack/python-manilaclient master: Fix manilclient.v2 import error
*** tellesnobrega has quit IRC14:33
openstackgerritMerged openstack/manila master: Enable IPv6 in manila(documentation)
openstackgerritOpenStack Proposal Bot proposed openstack/manila master: Updated from global requirements
*** harlowja has joined #openstack-manila14:51
openstackgerritTom Barron proposed openstack/manila-ui master: DO NOT MERGE
tbarrontosky: dunno if it will work but I'm trying with Depends-On: <yourchange>14:54
toskytbarron: I was going to ask the same, but see #openstack-release14:55
toskyjobs use on the released versions of the clients (not from git HEAD)14:55
*** kaisers has joined #openstack-manila14:57
*** Yogi1 has quit IRC15:01
toskytbarron: please change to depend on
toskyor I can do it15:08
toskyit should do the trick15:08
tbarrontosky: please just do it as my attention is diverted atm15:09
toskytbarron: sure :)15:09
* tbarron has been enjoying having raissa work in manila, maybe he can entice tosky over here too15:09
tosky>.> I'm fixing Sahara and this is like
tbarronwe just have to keep breaking sahara to achieve that goal15:10
raissatosky: welcome back to work eheh15:11
toskyeh :D15:11
openstackgerritLuigi Toscano proposed openstack/manila-ui master: DO NOT MERGE
*** dsariel has quit IRC15:14
*** tellesnobrega has joined #openstack-manila15:15
*** kaisers has quit IRC15:26
*** tellesnobrega has quit IRC15:26
*** kaisers has joined #openstack-manila15:26
tbarronganso: bswartz I think we should merge even if15:27
tbarronwe may need temporary fixes in manila-ui and sahara-ui (TBD whether we will)15:28
tbarronb/c it would be good to get the root cause fix in before stable/pike is cut (imminent)15:29
raissais that the right link?15:31
raissaok I guess it's
*** kaisers has quit IRC15:38
bswartztbarron: wrong link?15:43
tbarronraissa: thanks - bswartz 488427 - I truncated15:43
bswartzthanks raissa15:43
toskybswartz: see also the discussion on #openstack-release (you may see a ping), maybe a new release could be tagged for pike15:44
toskywith that fix15:44
toskyotherwise any user of 'import manilaclient.v2' should add that dependency15:44
bswartzI have no problem tagging another version15:48
bswartzit seems like a backport after stable/pike gets cut would require less rule bending15:49
bswartzI'm not clear on how bad the problem is15:49
toskytry to run 'import manilaclient.v2'15:50
toskyit explodes because ipaddress is missing15:50
toskyI would say "bad"15:51
*** harlowja has quit IRC15:51
*** jmlowe has quit IRC15:51
toskyalso because it's blocking pike3 for sahara-dashboard15:51
tbarroncurrent consumers of manilaclient being manila-ui and sahara-dashboard at least15:51
bswartzI don't understand how testing didn't catch this15:52
bswartzif it's that bad15:52
toskywhich uses manilaclient; I know that we could add the dependency, but according ttx on #openstack-release a new client can be easily tagged if the operation is done today15:52
toskyehm, I don't know why it was not catched15:52
bswartzpushing a new tag is easy, but I'd more interested in how we avoid the same problem in the future15:52
toskysure, it makes sense15:53
tbarronbswartz: i guess until the client was released ui and sahara dashboard jobs in gate didn't see the changes that introduced the need for the requirement15:54
toskyyes, but that's one part of it15:55
toskymanilaclient/v2/ ipaddress15:55
bswartztbarron: so there needs to be some kind of job in the manilaclient project that detects this condition15:55
toskybut no ipaddress in requirements; maybe it's covered elsewhere15:55
tbarronwe didn't need ipaddress explicitly in requirements in manila or in python-manilaclient itself15:56
toskyuhm, also manila.git imports ipaddress in various places, but I don't see in the requirements15:56
* tosky puzzled15:56
toskymaybe it was installed on the test images by default?15:56
tbarroni'm thinking maybe it was transitively (implicitly) required15:57
toskyso most likely a patch for manila is needed too15:58
tbarronit is present in my .tox/py27/lib/... and 'pip list/freeze' shows it in those components when I activate a .tox venv15:58
toskyand if you create a new py27 venv?16:00
tbarrontosky: probably, but not urgently I guess and manila server isn't imported16:01
tbarrontosky: I did 'rm -rf .tox' and rebuilt before looking16:01
tbarrons/I guess and/I guess since/16:02
toskytbarron: finds it in test-requirements.txt for oslo.serialization16:07
bswartztbarron: how confident are we that the patch moving through the gate is the right one?16:07
bswartzI saw the discussion in #release about various mechanisms to test it16:08
tbarronbswartz: it matches global-requirements.txt so it seems correct16:11
tbarronI think the question was whether if we can't get it in the release we need to also patch the consumers16:11
*** jmlowe has joined #openstack-manila16:12
toskybswartz: the various mechanisms don't work for non-devstack tests, like unit tests16:13
tbarronyeah, looks like dhellman's 3-way Depends-On trick doesn't work for this16:14
toskybut testing locally the failure in sahara-dashboard are gone16:15
tbarronbswartz: I don't think we know how to test this except that when the same requirement is added to the consumer it fixes the consumer.16:16
tbarrontosky: were you able to make sahara-dashboard Depend-On: the candidate python-manilaclient itself?16:17
toskytbarron: yes and no; I did, but it's pointless for unit tests16:17
*** dustins has quit IRC16:26
bswartzzuul says 0 minutes16:32
* bswartz waits for it.....16:32
openstackgerritMerged openstack/python-manilaclient master: Fix manilclient.v2 import error
*** kaisers has joined #openstack-manila16:34
openstackgerritOpenStack Proposal Bot proposed openstack/python-manilaclient master: Updated from global requirements
tbarron thanks tosky!16:36
toskynow I only hope that it was not the only blocker for sahara-dashboard; I get one last error but I hope it's my setup16:36
tbarrontosky: fingers crossed16:37
bswartztbarron: the bot doesn't like ltoscano's whitespace choices ^16:40
*** openstackstatus has quit IRC16:41
*** openstack has joined #openstack-manila16:45
tbarronthat would be handy16:47
toskyI don't think that gerrit has that information16:47
toskyit's in the launchpad profile most likely16:47
bswartzyeah launchpad is usually where I go to hunt down unknown names16:52
bswartzit would be nice if gerrit allowed you to add an IRC nick to your account16:52
*** catinthe_ has joined #openstack-manila16:54
*** harlowja has joined #openstack-manila16:54
*** harlowja_ has joined #openstack-manila16:55
*** catintheroof has quit IRC16:58
*** harlowja has quit IRC16:59
*** tellesnobrega has joined #openstack-manila17:00
*** kaisers has quit IRC17:02
openstackgerritBoris Pavlovic proposed openstack/manila master: Revert "Updated from global requirements"
*** kaisers has joined #openstack-manila17:04
-openstackstatus- NOTICE: The Gerrit service on is being taken offline for roughly 5 minutes to perform a database backup and reconfiguration17:12
*** dustins has joined #openstack-manila17:16
*** kaisers has quit IRC17:17
*** kaisers has joined #openstack-manila17:25
*** tosky has quit IRC17:31
*** tellesnobrega has quit IRC17:31
*** openstackgerrit has quit IRC17:33
*** kaisers has quit IRC17:45
*** kaisers has joined #openstack-manila17:45
*** kaisers has quit IRC17:49
*** catinthe_ has quit IRC18:08
*** catintheroof has joined #openstack-manila18:10
*** openstackstatus has quit IRC18:11
*** openstack has joined #openstack-manila18:15
bswartzganso: ping18:35
bswartzganso: I think I might have found a bug in mountable snapshots for LVM18:35
*** kaisers has joined #openstack-manila18:37
*** mayur__ has joined #openstack-manila18:41
*** openstackstatus has quit IRC18:42
*** openstack has joined #openstack-manila18:45
*** openstackgerrit has joined #openstack-manila18:51
openstackgerritBen Swartzlander proposed openstack/manila master: Change the way snapshots are exported in LVM
*** mayur__ has quit IRC18:54
openstackgerritBen Swartzlander proposed openstack/manila master: Changes to allow manila-test-image to work with scenario tests
*** chenying__ has quit IRC19:45
*** chenying__ has joined #openstack-manila19:47
*** cknight has quit IRC20:07
*** catintheroof has quit IRC20:08
gansobswartz: pong20:13
bswartzganso: !20:13
gansobswartz: gerrit is down20:13
gansobswartz: oh wait, it is not20:14
bswartzganso: so I did an experiment --
bswartzbut I wanted to ask you about this20:14
bswartzIt's not clear that there is a bug but I saw a test failure and that snapshot code made me suspicious20:14
bswartzI want to verify that there is in fact a bug before actually pushing for this fix20:15
gansobswartz: do you have a trace of the failure?20:16
bswartzin that test job 1 test failed, and upon inspection the reason was that snapshot creation failed during the "mount" command20:17
gansobswartz: no error in m-shr
bswartzthe "mount" command returned status 32 (device in use or busy)20:17
bswartzganso: grep for error, the button is broken20:17
gansobswartz: ah20:18
gansobswartz: you cannot consistently reproduce the problem, right?20:18
bswartzno I just saw it this one time20:19
bswartzbut people have been complaining about flaky lvm jobs lately20:19
gansobswartz: AFAIK, changing the order as you did in your patch does nothing20:19
bswartzso I'm digging into each and every LVM failure I see to find the root cause20:19
gansobswartz: the problem is that sometimes the mount operation takes a long time20:19
gansobswartz: sometimes it doesn't20:19
gansobswartz: tommylikehu invested a lot of time investigating this stuff20:20
bswartzso the issue with the existing code is that it sets up a share on the wrong directory20:20
bswartzit creates a CIFS share for an empty directory, then mounts the LV into that directory after the share exists20:20
bswartzit's unlikely that Samba does the right thing (need to confirm that though)20:20
bswartzin any case, do you see any downsides to that change I proposed?20:21
gansobswartz: it seems you are creating the export before mounting the LVM20:21
bswartzthe existing code creates the export then mounts20:21
bswartzmy proposal is to mount then create the export20:21
*** boris_42 has joined #openstack-manila20:21
boris_42vponomaryov bswartz hi there20:22
gansobswartz: oh right, sorry, was reading the other way around20:22
bswartzintuitively, mounting first seems like the right thing to do20:22
boris_42Rally Manila job start failing few days ago20:22
gansobswartz: just a sec20:22
bswartzbut since you were co-author on the patch that added these lines I wanted to ask you20:22
boris_42did you guys change anything?)20:22
bswartzboris_42: yes you might have hit
openstackLaunchpad bug 1707194 in python-manilaclient "test: gate: Failure: ImportError (cannot import name availability_zones)" [Critical,In progress] - Assigned to Tom Barron (tpb)20:23
boris_42@bswartz so you are already working on this issue?)20:23
gansobswartz: I think your change is correct20:23
boris_42bswartz just making sure that things are undercontrol =)20:23
bswartzboris_42: feature freeze was yesterday and a bunch of code merged this week20:23
gansobswartz: but I am unsure of why the current code is flaky20:23
bswartzwe had some critical bugs20:23
boris_42@bswartz that's the evil thing about feature freeze =)20:23
boris_42creates chaos =)20:23
bswartzbut in theory everything is fixed now20:24
bswartzif we continue to see problems we will investigate and fix them20:24
bswartzganso: yes it's hard to understand how the current code would behave nondeterministically -- it seems pretty straightforward20:24
bswartzI wanted to do a good scenario test of that20:24
bswartzcreate a share with some data, then snapshot it and mount the snapshot (with CIFS)20:25
boris_42bswartz: I'll recheck the job20:25
bswartzboris_42: what was the time of the earliest failure?20:25
boris_42@bswartz something about 2-3 days ago20:26
boris_42@bswartz I resubmit emtpy change patch to see whatever gates are passing or not20:26
boris_42@bswartz that is going to give us result in about 20 minutes20:26
bswartzboris_42: ty20:26
*** jmlowe has quit IRC20:39
boris_42@bswartz its' still failing with the same errors20:44
boris_42  File "/opt/stack/new/python-manilaclient/manilaclient/v2/", line 328, in list20:45
boris_42    search_opts.pop("export_location", None)20:45
boris_42AttributeError: 'NoneType' object has no attribute 'pop20:45
boris_42@bswartz seems like it's some typo in client during the recent work on shares20:48
*** kaisers has quit IRC20:51
gansobswartz: yes. But logically it seems incorrect. It sure slipped under our radars20:55
*** kaisers has joined #openstack-manila20:58
gouthamrboris_42: still around?21:11
gouthamris there an LP for the bug you found just now?21:12
boris_42*big brother is watching you*21:12
boris_42so no bug I just saw that manila gates started failing21:12
boris_42and we didn't intorduce (from Rally side anything that may cause that)21:12
gouthamrboris_42: yes... true, i'll file one. thanks for noticing!21:13
openstackgerritMerged openstack/python-manilaclient master: Updated from global requirements
openstackLaunchpad bug 1707303 in python-manilaclient "Manilaclient errors on share list for API versions <=2.34" [Undecided,New]21:14
boris_42@gouthamr thanks21:15
openstackgerritGoutham Pacha Ravi proposed openstack/python-manilaclient master: Fix list command when not given search_opts
*** gouthamr has quit IRC21:24
*** dustins has quit IRC21:33
*** gouthamr has joined #openstack-manila21:36
boris_42@gouthamr btw I did this change just to check whatever your fix works or not21:45
boris_42@gouthamr and thanks for quick fix21:45
gouthamrboris_42: np.. but wouldn't the job still pull down python-manilaclient>=1.12.0,<=1.16.021:50
boris_42gouthamr: nope it uses upstream clients21:50
boris_42gouthamr: it's dark magic of CI jobs21:50
boris_42but it works=)21:50
gouthamrhaha.. dark magic really, finding all these bugs. :)21:51
*** kaisers has quit IRC21:52
boris_42@gouthamr both jobs passed this works works as expected ;)22:02
gouthamrboris_42: thanks, we'll merge it22:02
*** jmlowe has joined #openstack-manila22:18
boris_42gouthamr: btw maybe it makes sense to all rally jobs to client repo as well?22:19
gouthamrboris_42: absolutely... they're not long running jobs..22:22
boris_42@gouthamr ya and that is going to catch some issues as well22:22
*** kaisers has joined #openstack-manila22:23
gouthamrboris_42: adding them to the check queue would suffice?22:26
gouthamri don't think we need them in the gate queue, vponomaryov may have other ideas..22:27
*** kaisers has quit IRC22:27
boris_42@gouthamr ya just in check22:32
boris_42@gouthamr as non voting22:32
gouthamrboris_42: these two jobs are non-voting in general, across all projects that have them enabled..22:33
boris_42@gouthamr ya22:34
gouthamrboris_42: awesome. thanks for pointing this out!22:34
*** ganso has quit IRC23:23
openstackgerritMerged openstack/manila master: Updated from global requirements

Generated by 2.15.3 by Marius Gedminas - find it at!