Friday, 2018-02-16

« Thursday, 2018-02-15 Index Saturday, 2018-02-17 »
*** salv-orlando has quit IRC00:06
tommylikehuhey pooja_jadhav, sorry for late response, my point here is that it would be great if we can support boolean value with extra spaces in all of the APIs after json schema changes.00:07
*** salv-orlando has joined #openstack-cinder00:07
*** salv-orlando has quit IRC00:12
*** AlexeyAbashkin has joined #openstack-cinder00:13
*** DHE has left #openstack-cinder00:13
*** AlexeyAbashkin has quit IRC00:18
*** gman-tx has quit IRC00:18
*** Dinesh_Bhor has joined #openstack-cinder00:42
*** Dinesh_Bhor has quit IRC00:42
*** andreas_s has joined #openstack-cinder00:45
*** andreas_s has quit IRC00:49
*** gman-tx has joined #openstack-cinder00:54
*** felipemonteiro_ has quit IRC01:03
*** lhx_ has joined #openstack-cinder01:05
*** salv-orlando has joined #openstack-cinder01:08
*** gman-tx has quit IRC01:11
*** salv-orlando has quit IRC01:12
*** AlexeyAbashkin has joined #openstack-cinder01:13
*** AlexeyAbashkin has quit IRC01:18
openstackgerritMerged openstack/cinder master: Admin guide documentation for volume multiattach  https://review.openstack.org/54004701:51
*** salv-orlando has joined #openstack-cinder02:08
*** salv-orlando has quit IRC02:13
*** mriedem1 has joined #openstack-cinder02:29
*** mriedem has quit IRC02:29
*** mriedem1 is now known as mriedem02:29
*** lhx_ has quit IRC02:46
*** kaisers__ has joined #openstack-cinder02:57
*** dave-mccowan has joined #openstack-cinder03:01
*** kaisers_ has quit IRC03:02
*** mriedem has quit IRC03:07
*** salv-orlando has joined #openstack-cinder03:09
*** salv-orlando has quit IRC03:14
*** dave-mcc_ has joined #openstack-cinder03:15
*** dave-mccowan has quit IRC03:15
*** harlowja_ has quit IRC03:23
openstackgerritMohammed Naser proposed openstack/cinder master: Add throughput limits that scale per-GB  https://review.openstack.org/54520103:23
*** vivsoni has joined #openstack-cinder03:24
mnaser^ how do people feel about backporting that... i reaaaaally don't want to end up maintaining our own branch, it's pretty important for us to have and it is fairly low impact03:29
*** abhishekk has joined #openstack-cinder03:39
*** dave-mcc_ has quit IRC03:41
*** gkadam has joined #openstack-cinder03:42
*** gman-tx has joined #openstack-cinder03:44
*** gman-tx has quit IRC03:52
*** andreas_s has joined #openstack-cinder04:13
*** andreas_s has quit IRC04:17
*** salv-orlando has joined #openstack-cinder04:23
*** psachin has joined #openstack-cinder04:26
*** salv-orlando has quit IRC04:29
*** lpetrut has joined #openstack-cinder04:36
*** harlowja has joined #openstack-cinder04:46
*** dalgaaf has quit IRC04:55
*** dalgaaf has joined #openstack-cinder04:56
*** harlowja has quit IRC04:59
*** mdbooth has quit IRC05:13
*** lpetrut has quit IRC05:13
*** lpetrut has joined #openstack-cinder05:19
*** harlowja has joined #openstack-cinder05:19
*** salv-orlando has joined #openstack-cinder05:25
*** dobson has quit IRC05:25
*** harlowja has quit IRC05:27
*** salv-orlando has quit IRC05:29
*** lpetrut has quit IRC05:35
*** dobson has joined #openstack-cinder05:37
*** hoonetorg has quit IRC05:40
*** hoonetorg has joined #openstack-cinder05:53
*** kaisers__ has quit IRC05:54
*** kaisers_ has joined #openstack-cinder06:02
*** pgadiya has joined #openstack-cinder06:08
*** pgadiya has quit IRC06:17
openstackgerritIbad Khan proposed openstack/cinder master: QNAP Drivers - Move from httplib to requests  https://review.openstack.org/53823706:19
*** threestrands has quit IRC06:20
*** salv-orlando has joined #openstack-cinder06:25
*** openstackstatus has quit IRC06:27
*** openstack has joined #openstack-cinder06:32
*** ChanServ sets mode: +o openstack06:32
*** salv-orlando has joined #openstack-cinder06:32
openstackgerritVivek Soni proposed openstack/cinder master: 3PAR: Monitor task of promoting a virtual copy  https://review.openstack.org/54521406:55
*** andreas_s has joined #openstack-cinder06:57
*** andreas_s has quit IRC07:01
*** gkadam has quit IRC07:04
*** gkadam has joined #openstack-cinder07:10
*** pgadiya has quit IRC07:18
*** threestrands has joined #openstack-cinder07:22
*** threestrands has quit IRC07:22
*** threestrands has joined #openstack-cinder07:23
*** threestrands has joined #openstack-cinder07:23
*** AlexeyAbashkin has joined #openstack-cinder07:26
*** pgadiya has joined #openstack-cinder07:32
*** gkadam has quit IRC07:33
*** threestrands has quit IRC07:34
*** rcernin has quit IRC07:41
*** lpetrut has joined #openstack-cinder07:44
*** andreas_s has joined #openstack-cinder08:10
*** alexchadin has joined #openstack-cinder08:13
*** alexchadin has quit IRC08:24
*** alexchadin has joined #openstack-cinder08:25
*** tesseract has joined #openstack-cinder08:27
*** salv-orlando has quit IRC08:32
*** salv-orlando has joined #openstack-cinder08:32
*** e0ne has joined #openstack-cinder08:35
*** salv-orlando has quit IRC08:37
*** salv-orlando has joined #openstack-cinder08:45
*** david-lyle has quit IRC08:46
*** e0ne has quit IRC08:54
*** yangyapeng has joined #openstack-cinder08:57
*** salv-orlando has quit IRC09:04
*** salv-orlando has joined #openstack-cinder09:05
*** salv-orlando has quit IRC09:09
*** belmoreira has joined #openstack-cinder09:11
*** pcaruana has joined #openstack-cinder09:11
*** frickler has joined #openstack-cinder09:19
fricklerare there any docs available for how to deploy https://specs.openstack.org/openstack/cinder-specs/specs/mitaka/cinder-volume-active-active-support.html ?09:25
*** stephenfin is now known as finucannot09:35
*** links has joined #openstack-cinder09:47
*** links has quit IRC09:49
*** ganso has joined #openstack-cinder09:56
*** links has joined #openstack-cinder09:58
*** abhishekk has quit IRC09:58
*** alexchadin has quit IRC10:01
*** links has quit IRC10:12
*** openstackgerrit has quit IRC10:18
*** luizbag has joined #openstack-cinder10:23
*** dosaboy has quit IRC10:33
*** stakeda has quit IRC10:33
*** e0ne has joined #openstack-cinder10:34
*** dosaboy has joined #openstack-cinder10:38
vivsoniTeam, please review https://review.openstack.org/#/c/545214/10:49
zigoHi guys. I have a bunch of failed tests when switching Cinder to Python 3.10:53
zigoNamely: cinder.tests.unit.api.contrib.test_backups.BackupsAPITestCase.test_restore_backup_with_body_KeyError10:54
zigocinder.tests.unit.api.test_api_validation.PatternPropertiesTestCase.test_validate_patternProperties_fails10:54
zigocinder.tests.unit.volume.drivers.dell_emc.vmax.test_vmax.VMAXCommonReplicationTest.test_replicate_group10:54
zigoplus lots of failed tests in qnap and netapp drivers.10:54
*** alexchadin has joined #openstack-cinder10:54
zigoI filed a few bugs for them.11:07
*** alexchadin has quit IRC11:10
*** e0ne has quit IRC11:12
*** e0ne has joined #openstack-cinder11:13
*** alexchadin has joined #openstack-cinder11:22
*** tpsilva has joined #openstack-cinder11:23
*** armaan has joined #openstack-cinder11:27
*** belmoreira has quit IRC11:28
*** armaan has quit IRC11:35
*** AlexeyAbashkin has quit IRC11:37
*** AlexeyAbashkin has joined #openstack-cinder11:40
*** alexchadin has quit IRC11:41
*** alexchadin has joined #openstack-cinder11:42
*** pgadiya has quit IRC11:49
zigohttps://bugs.launchpad.net/bugs/1749927 https://bugs.launchpad.net/bugs/1749929 https://bugs.launchpad.net/bugs/1749930 https://bugs.launchpad.net/bugs/1749932 https://bugs.launchpad.net/bugs/174993411:53
openstackLaunchpad bug 1749927 in Cinder "Unit test test_restore_backup_with_body_KeyError fails under Python 3.6" [Undecided,New]11:53
openstackLaunchpad bug 1749929 in Cinder "Unit test test_validate_patternProperties_fails fails under Python 3.6" [Undecided,New]11:53
openstackLaunchpad bug 1749930 in Cinder "Unit test VMAXCommonReplicationTest.test_replicate_group fails under Python 3.6" [Undecided,New]11:53
openstackLaunchpad bug 1749932 in Cinder "4 of netapp unit tests failing in Python 3.6" [Undecided,New]11:53
openstackLaunchpad bug 1749934 in Cinder "46 failed unit tests for qnap when running in Python 3.6" [Undecided,New]11:53
*** links has joined #openstack-cinder12:03
e0nezigo: we've got only python 3.5 support now12:06
zigoe0ne: Well, that's the thing, Debian Sid uses Python 3.6, Gentoo has 3.6 too, and Fedora will do 3.6 for Rocky. So these bugs must be addressed at some point.12:07
e0nezigo: do you volunteer to help us with fixes?12:08
zigoe0ne: I can attempt to help someone that will try to fix, but I cannot say I will fix the issues myself. I can't do that with all of the 421 OpenStack packages I maintain in Debian, you know.12:09
e0nezigo: I agree that it would be good to fix them, but we don't test with 3.6 on gates12:09
e0nehonestly, we even don't have python3 in setup.cfg now:(12:10
zigoe0ne: I'm ok with no gating yet, as long as the issues are taken care of. IMO, the most important ones to fix are the first 2 in the list above, since they are not related to a particular driver.12:10
*** salv-orlando has joined #openstack-cinder12:10
zigoWell, this better change ASAP then, because everyone is moving toward Py 3.12:11
zigoUbuntu will do so for Rocky, and so is RDO.12:11
*** armaan has joined #openstack-cinder12:13
*** psachin has quit IRC12:14
*** psachin has joined #openstack-cinder12:25
*** elmaciej has joined #openstack-cinder12:26
*** elmaciej_ has joined #openstack-cinder12:28
*** elmaciej has quit IRC12:32
*** belmoreira has joined #openstack-cinder12:43
*** dave-mccowan has joined #openstack-cinder13:08
*** nicolasbock has joined #openstack-cinder13:10
*** nicolasbock has quit IRC13:10
*** r-daneel has joined #openstack-cinder13:14
*** armaan has quit IRC13:19
*** zhongjun has quit IRC13:20
*** openstackgerrit has joined #openstack-cinder13:22
openstackgerritMohammed Naser proposed openstack/cinder master: Add throughput limits that scale per-GB  https://review.openstack.org/54520113:22
*** armaan has joined #openstack-cinder13:25
*** dave-mccowan has quit IRC13:31
*** armaan has quit IRC13:32
*** abishop has joined #openstack-cinder13:36
*** psachin has quit IRC13:42
*** gman-tx has joined #openstack-cinder13:43
*** eharney has quit IRC13:43
*** gman-tx has quit IRC13:47
*** dustins has joined #openstack-cinder13:52
*** liverpooler has joined #openstack-cinder13:55
*** gman-tx has joined #openstack-cinder13:57
*** pchavva has joined #openstack-cinder13:59
*** mriedem has joined #openstack-cinder14:03
*** dave-mccowan has joined #openstack-cinder14:05
lpetrutHi, we have a bug fix on the SMB driver, if anyone could please take a look (it already has a +2): https://review.openstack.org/#/c/539593/ Two other patches have been +w-ed but depend on it.14:18
smcginnisWe target python 2.7 and python 3.5. That's the offical OpenStack-wide stance.14:21
*** andreas_s has quit IRC14:24
*** andreas_s has joined #openstack-cinder14:25
*** eharney has joined #openstack-cinder14:26
*** felipemonteiro_ has joined #openstack-cinder14:30
mriedemwhat does this mean exactly? https://docs.openstack.org/cinder/latest/contributor/api_microversion_history.html#id4714:31
mriedemi've created a multiattach capable type following https://docs.openstack.org/cinder/latest/admin/blockstorage-volume-multiattach.html#multiattach-volume-type14:31
mriedemis 3.50 only applied when i go to create a volume using that type?14:31
mriedemi'm able to create the type and set the extra spec at 3.0 no problem14:32
*** elmaciej_ has quit IRC14:32
*** elmaciej has joined #openstack-cinder14:33
smcginnisHmm, that should probably be expanded a little.14:33
*** felipemonteiro__ has joined #openstack-cinder14:34
*** andreas_s has quit IRC14:35
mriedemusing 3.50 doesn't seem to make a difference when i create the volume, it always fails during scheduling saying there is no backend that supports it...14:36
mriedemFeb 16 14:35:59 queens cinder-scheduler[1556]: ERROR cinder.scheduler.flows.create_volume [None req-e5036fa1-c54f-4410-a57b-c2cc5f96fd53 demo None] Failed to run task cinder.scheduler.flows.create_volume.ScheduleCreateVolumeTask;volume:create: No valid backend was found. No weighed backends available: NoValidBackend: No valid backend was found. No weighed backends available14:37
*** alexchadin has quit IRC14:37
smcginnisLooking at the code, I actually am not sure why 3.50 was added.14:37
*** felipemonteiro_ has quit IRC14:37
mriedemi created this devstack env on the 9th so lvm should support it14:37
mriedemi'm actually trying to test that the policy rule works, such that if i set the policy to disable multiattach and try to create a volume as a non-admin, it should fail, but that doesn't seem to happen14:38
mriedemlbragstad: should /etc/cinder/policy.yaml be picked up automatically?14:38
mriedemstack@queens:~/devstack$ cat /etc/cinder/policy.yaml14:38
mriedem'volume:multiattach': 'rule:admin'14:38
smcginnisI'm looking at the actual patch that added 3.50 and I don't see the policy or the mv being checked.14:38
mriedemthe microversion might have just been for discoverability14:39
smcginnisSo if 3.50 or greater is not specified, should we ignore that in the volume type?14:39
*** elmaciej has quit IRC14:39
smcginnisIt's not really an API change to me.14:39
smcginnisOh, true.14:39
smcginnisBut should we ignore it if mv is < 3.50?14:39
smcginnisAnd if policy is not allowed, should we just not set multiattach, or should we actually fail the op?14:40
*** elmaciej has joined #openstack-cinder14:40
mriedemi'm just trying to create a demo with the type since that's the recommended way to do this, but i could just use the multiattach flag on the volume too14:40
mriedemi need to figure out why the type fails during scheduling though14:40
mriedemi just did14:40
mriedem$ openstack volume type create --property multiattach="<is> True" multiattach14:40
mriedemis that enough?14:40
mriedemwill it default to lvm otherwise?14:41
smcginnisI thought that was it.14:41
mriedemoh hmm14:41
mriedemstack@queens:~/devstack$ openstack volume type show multiattach14:41
mriedem+--------------------+--------------------------------------+14:41
mriedem| Field              | Value                                |14:41
mriedem+--------------------+--------------------------------------+14:41
mriedem| access_project_ids | None                                 |14:41
mriedem| description        | None                                 |14:41
mriedem| id                 | a17cff38-51b4-4250-80f3-cf8954547859 |14:41
mriedem| is_public          | True                                 |14:41
mriedem| name               | multiattach                          |14:41
mriedem| properties         |                                      |14:41
mriedem+--------------------+--------------------------------------+14:41
mriedemthe property doesn't seem to be set there14:41
smcginnisYou can specify backend name to make it explicit, but it _should_ be able to select it based on lvm being the only capable backend.14:41
smcginnisHmmm14:41
*** dansmith is now known as superdan14:43
*** elmaciej_ has joined #openstack-cinder14:43
lbragstadmriedem by picked up automatically?14:44
lbragstaddo you mean it has overrides in it and they aren't registering?14:44
mriedemlbragstad: yeah i just wrote one rule to /etc/cinder/policy.yaml14:44
mriedemand it didn't seem to get validated14:44
mriedembut now i'm maybe hitting something different14:44
lbragstadmriedem yeah - as long as the configuration file specifies the correct policy file, that should be automatic14:44
openstackgerritMerged openstack/cinder master: 3PAR: Monitor task of promoting a virtual copy  https://review.openstack.org/54521414:45
mriedemwhich config? cinder.conf?14:45
lbragstadyeah14:45
mriedemsmcginnis: this time the type looks OK http://paste.openstack.org/show/674484/14:45
mriedemweird14:45
smcginnisI blame the user.14:45
lbragstadoslo.policy passes options through each services configuration file for common policy stuff14:45
mriedemlbragstad: ok cinder.conf might default to policy.json14:45
* lbragstad digs14:45
mriedemyup #policy_file = policy.json14:45
lbragstadhttps://github.com/openstack/oslo.policy/blob/master/oslo_policy/opts.py14:46
mriedemok i figured policy.yaml was too futuristic14:46
lbragstadnice14:46
*** elmaciej has quit IRC14:46
lbragstadthat yaml is some pretty crazy stuff14:46
lbragstadif you set the path to point to policy.yaml, it should register that one override automatically in cinder's policies14:47
mriedemyeah for the sake of this demo i'll just use policy.json for the defaults14:47
smcginnisWe are at least checking policy: https://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L29214:49
smcginnisBut not mv.14:49
mriedemyeah testing that now14:49
mriedemaha14:50
mriedemstack@queens:~$ openstack volume create --image cirros-0.3.5-x86_64-disk --size 1 --type multiattach cirros-multiattach-volume14:50
mriedemPolicy doesn't allow volume:multiattach to be performed. (HTTP 403) (Request-ID: req-e98ce62f-69b7-44c3-a60d-f711af5d8241)14:50
mriedemand that was w/o 3.5014:50
smcginnisLooks like we need to add a check.14:50
mriedemfor mv?14:50
mriedemlike i said, i think it's just there for discoverability14:51
mriedembut you'd have to confirm with jgriffith14:51
smcginnisI guess I could see three options..14:51
smcginnis1) Must specify 3.50 or later or it gets ignored and is not set to multiattach14:51
smcginnis2) Must specify 3.50 or later or it errors out if multiattach is specified14:52
smcginnis3) It's just a programmatic way to check and we don't give a rats ass what the request comes in with.14:52
*** jistr is now known as jistr|mtg14:52
smcginnisI guess we've gone the option 3 route with how things are right now.14:52
mriedemyeah14:52
smcginnisWhich honestly is just fine with me.14:52
mriedemat this point it'd be bad to go back on that probably14:53
mriedemdid you guys do an rc2?14:53
smcginnisOne planned for Monday.14:53
smcginnisBut I don't really think we should change it, thinking about it.14:53
mriedem(2) gets weird with the existing multiattach flag on the volume that's been around since forever14:53
smcginnisCould be confusing, especially to end users that don't have visibility into extra specs and such.14:53
smcginnismriedem: Right, that's worse than handling the new style.14:54
*** andreas_s has joined #openstack-cinder14:54
mriedemdefault policy probably shows the type extra specs (that's the way it is in nova i think), but i agree that needing to look at extra specs as a user sucks14:54
mriedemok so i'll keep going with this14:54
jungleboyjmriedem:  ++14:54
*** salv-orl_ has joined #openstack-cinder14:55
smcginnisOh good, my last concern was that we do not check policy on retype.14:55
smcginnisBut that is covered.14:55
smcginnishttps://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L165614:55
smcginnisAlthough, in that case it will fail if going from a multiattach type to a non-multiattach type.14:56
smcginnisWhich seems a little odd to me.14:56
mriedemthat should only be for in-use volumes i think14:56
smcginnisNo, that's outside of the if block14:56
jungleboyjsmcginnis:  That is a bug then.14:57
mriedemwe don't want to retype from / to multiattach when a volume is attached b/c that triggeres swap volume in nova which could be bad14:57
smcginnisIt's if the source volume and the target volume have different multiattach settings, so you could be going from a capable one to an incapable one, and it would fail saying your not allowed to do the thing that you're trying not to do.14:57
mriedemin ^ it checks the volume status and if it's not available it would fail before the policy check14:57
smcginnismriedem: It's outside of that if block.14:57
*** mdbooth has joined #openstack-cinder14:57
smcginnisThis is where {} comes in handy.14:58
smcginnisOr blank lines for readability.14:58
*** r-daneel has quit IRC14:58
*** salv-orlando has quit IRC14:58
*** andreas_s has quit IRC14:58
mriedemmultiattach_policy is admin_or_owner; so you should be able to retype to/from multiattach capability as long as the volume is available14:59
mriedemwhich is what that code is doing right?14:59
smcginnishttps://github.com/openstack/cinder/blob/master/cinder/volume/api.py#L165614:59
mriedemright, i don't see the issue14:59
mriedemis it that, i could be going from multiattach=True to multiattach=False and fail the policy check?15:00
mriedemeven though i'm going to multiattach=False?15:00
*** pooja_jadhav has quit IRC15:00
openstackgerritSean McGinnis proposed openstack/cinder master: Only enfore multiattach policy for retype target  https://review.openstack.org/54535715:00
smcginnismriedem: Correct.15:00
openstackgerritSean McGinnis proposed openstack/cinder master: Only enforce multiattach policy for retype target  https://review.openstack.org/54535715:01
ildikovmriedem: so the microversion was just for discoverability15:02
ildikovmriedem: we didn't have any enforcement like how we do in the Nova API15:02
mriedemsmcginnis: logic looks good in that patch15:02
mriedemback to figuring out why i can't actually create a volume using this type...15:03
smcginnismriedem: Cool. I figured it would make more sense after seeing that.15:03
mriedemhmm, yeah i continue to fail in scheduling15:03
*** andreas_s has joined #openstack-cinder15:04
ildikovmriedem: does it fail with both setting the flag and the volume_type method?15:06
*** felipemonteiro__ has quit IRC15:06
*** felipemonteiro_ has joined #openstack-cinder15:06
mriedemi haven't tried the flag15:06
mriedemhaving a hard time simply deleting volumes15:06
mriedemhttp://paste.openstack.org/show/674508/15:06
ildikovthat's weird15:07
mriedemno errors in the logs when trying to force delete that volume15:07
mriedembut it won't go away15:07
mriedemaha15:08
mriedemFeb 16 15:08:02 queens cinder-volume[1577]: ERROR cinder.utils [req-50b3a91d-56e4-4ec5-a89d-adc147d57ea4 req-7f8a6455-3e61-48e2-a152-3092221c9784 demo None] Volume driver LVMVolumeDriver not initialized15:08
*** andreas_s has quit IRC15:08
mriedemwhich is probably the reason for my schedluing failures15:09
jungleboyjsmcginnis:  Is that another one for RC2?15:09
ildikovmriedem: yeah, that explains it15:09
mriedemi guess this is preventing c-vol from starting15:10
mriedem ERROR cinder.volume.manager Stderr: u'  Volume group "stack-volumes-lvmdriver-1" not found\n  Cannot process volume group stack-volumes-lvmdriver-1\n'15:10
mriedemi don't know how that would be gone15:11
ildikovhmm, I don't remember seeing that before15:12
jungleboyjsmcginnis:  Merging this one for you:  https://review.openstack.org/#/c/545210/115:14
*** lbragstad has quit IRC15:14
openstackgerritSean McGinnis proposed openstack/cinder master: Expand on mv 3.50 description  https://review.openstack.org/54536115:14
*** jistr|mtg is now known as jistr15:15
*** lbragstad has joined #openstack-cinder15:16
openstackgerritMatt Riedemann proposed openstack/cinder master: Expand on mv 3.50 description  https://review.openstack.org/54536115:16
smcginnismriedem: :P15:17
smcginnisjungleboyj: ++15:17
jungleboyjEverything I know about reviewing I learned from mriedem15:17
jungleboyjDidn't you have another one that needed BP for Queens?15:17
mriedemrecreated the vg15:20
mriedemrestarted c-vol15:20
mriedemthings are up now15:20
mriedemfinally force deleted that volume (after changing the policy, yay)15:22
mriedembooya15:23
mriedemstack@queens:~$ openstack volume show -c name -c multiattach -c status cirros-multiattach-volume15:23
mriedem+-------------+---------------------------+15:23
mriedem| Field       | Value                     |15:23
mriedem+-------------+---------------------------+15:23
mriedem| multiattach | True                      |15:23
mriedem| name        | cirros-multiattach-volume |15:23
mriedem| status      | available                 |15:23
mriedem+-------------+---------------------------+15:23
jungleboyjsmcginnis:  You ok with the time I moved Cinder's picture to?15:25
*** felipemonteiro__ has joined #openstack-cinder15:26
smcginnisjungleboyj: I think that should work for me.15:26
jungleboyjOk and should minimize disruption.15:26
*** felipemonteiro_ has quit IRC15:29
*** lpetrut has quit IRC15:31
ildikovmriedem: \o/15:36
ildikovmriedem: are you recording all the obstacles for the demo too? :)15:36
*** andreas_s has joined #openstack-cinder15:36
mriedemildikov: just did, hit another one15:37
ildikov:/15:37
mriedemwe have novaclient capped at 9.1.1 in queens, and you need novaclient 10.1.0 to get 2.60 support for OSC since OSC uses novaclient15:37
mriedemi could just use nova CLI but wanted to only use OSC for the demo15:37
ildikovah, ok15:38
mriedemthe joys of being a user15:39
mriedemsomething i don't fart around with often enough15:39
ildikovyeah, I'm out of practice as well15:39
*** r-daneel has joined #openstack-cinder15:40
jungleboyjI have been a user lately and it is a difference experience.  :-)15:40
mriedemit's an easy way to find problems in all of the docs15:40
jungleboyj:-)15:40
*** e0ne_ has joined #openstack-cinder15:43
*** e0ne has quit IRC15:43
*** david-lyle has joined #openstack-cinder15:44
ildikovI know, when we have the Upstream training I always tell the students to install DevStack and start playing with it; it's guaranteed they will quickly find a bug for themselves to fix15:44
*** e0ne has joined #openstack-cinder15:45
*** swamireddy has quit IRC15:47
jungleboyj:-)15:47
*** e0ne_ has quit IRC15:47
mriedemheh, hit the latent nova bug where bfv fails in the api, you delete the server and the volume is left 'reserved'15:48
mriedemso you have to force detach it15:48
* mriedem goes to change default policy yet again15:48
mriedemthere is no force detach CLI anywhere?15:49
mriedemi can reset-state, but that doesn't cleanup the attachments on the volume right?15:51
ildikovaccording to my knowledge it does not15:52
jungleboyjmriedem:  That is true.15:52
mriedemalright, curling it up15:54
*** r-daneel has quit IRC15:55
*** csaikia has joined #openstack-cinder15:56
*** pcaruana has quit IRC15:57
mriedemhmm, what am i doing wrong here?15:59
mriedemcurl -d '{"os-force_detach": {}}' -H "accept: application/json" -H "x-auth-token: $token" http://199.204.45.19/volume/v3/e9d773beeef2435eb59f7c6eeaf685a9/volumes/126c8d4b-c582-484a-8c09-fe901a7dc17f/action15:59
mriedem{"badRequest": {"message": "There is no such action: None", "code": 400}}15:59
*** salv-orl_ has quit IRC16:00
*** salv-orlando has joined #openstack-cinder16:00
*** felipemonteiro__ has quit IRC16:02
*** r-daneel has joined #openstack-cinder16:03
*** salv-orlando has quit IRC16:05
geguileomriedem: I believe you are missing the connector information within the os-force_detach16:05
geguileodictionary16:06
mriedemgeguileo: it's optional16:06
mriedemaccording to the api-ref16:06
mriedemhttps://developer.openstack.org/api-ref/block-storage/v3/#force-detach-a-volume16:06
geguileooh, true, true16:06
mriedemit appears to not even get to the actual method16:08
*** andreas_s has quit IRC16:09
mriedemAHA16:09
mriedemFeb 16 16:08:49 queens devstack@c-api.service[1549]: DEBUG cinder.api.openstack.wsgi [None req-c7279a60-f7ba-4a11-98f2-8fa2b2ec281d demo demo] Unrecognized Content-Type provided in request {{(pid=1723) get_body /opt/stack/cinder/cinder/api/openstack/wsgi.py:724}}16:09
*** finucannot is now known as stephenfin16:10
geguileook, you are missing  -H "Content-Type: application/json"16:11
mriedemyeah i was using an accept header, needed content-type16:12
*** david-lyle has quit IRC16:15
*** slunkad_ has joined #openstack-cinder16:17
*** david-lyle has joined #openstack-cinder16:22
*** luizbag has quit IRC16:22
*** markvoelker has quit IRC16:23
erloneharney, hey, can you give a look on these when you have a chance? Added the depends-on and cherry-picked tags16:24
eharneyerlon: sure16:26
*** andreas_s has joined #openstack-cinder16:27
*** markvoelker_ has joined #openstack-cinder16:27
erloneharney, this is aquick one too: https://review.openstack.org/#/c/543912/16:28
* mriedem obviously doesn't curl much16:30
openstackgerritMatt Riedemann proposed openstack/cinder master: Provide a hint when performing a volume action can't find the method  https://review.openstack.org/54537816:32
*** aspiers has quit IRC16:33
*** AlexeyAbashkin has quit IRC16:33
*** imacdonn has quit IRC16:37
*** imacdonn has joined #openstack-cinder16:38
*** felipemonteiro has joined #openstack-cinder16:45
mriedemildikov: oh gdi,16:46
*** andreas_s has quit IRC16:46
mriedemmy devstack is using pike uca so the qemu version is too high for multiattach... :)16:46
mriedemtime to recreate devstack...16:46
*** aspiers has joined #openstack-cinder16:49
openstackgerritEric Harney proposed openstack/cinder master: Enable hash randomization in unit tests  https://review.openstack.org/46563616:50
openstackgerritEric Harney proposed openstack/cinder master: QNAP: Pass unit tests with hash randomization enabled  https://review.openstack.org/54538316:50
smcginniseharney: \o/16:50
eharneya few more off the list... more to go16:51
smcginnisIt's at least progress.16:51
smcginnisSounds like after 18.04 is released in April, or shortly thereafter, python 3 testing will move to 3.6.16:51
smcginnisSo... I guess we have some extra motivation to get that stuff moving.16:52
ildikovmriedem: I wanted to ask about that, but I was sure that if someone, you have the right env :)16:53
mriedemi created this devstack env last week before thinking about doinga demo16:53
mriedembut probably would have forgotten anyway..16:53
mriedemneed to create a multiattach-specific localrc file16:54
ildikovyeah, multi-attach is special :)16:54
mriedemmy ever expanding set of local.confs16:54
ildikovnever ending story16:55
smcginnisThat movie confused me so much as a kid.16:55
*** itlinux has joined #openstack-cinder16:55
mriedemi hadn't seen it until i was in my late 20s with my now wife,16:56
mriedemtotally didn't see what all of the hubbub was about16:56
ildikovsmcginnis: lol16:56
ildikovit confused me a little too TBH :)16:56
smcginnisThere's probably about a year and a half window in age where it's good an makes some sense. Otherwise it's either too confusing or you're too old. :)16:57
*** luizbag has joined #openstack-cinder16:58
*** r-daneel has quit IRC17:02
*** e0ne has quit IRC17:06
*** elmaciej_ has quit IRC17:07
SwansonI have the same opinion of A Clockwork Orange.17:08
*** felipemonteiro has quit IRC17:10
smcginnisNow that one I saw way too young and was definitely confused by. Only slightly better older after having read the book.17:11
*** luizbag has quit IRC17:11
*** luizbag has joined #openstack-cinder17:14
*** links has quit IRC17:18
*** itlinux has quit IRC17:22
*** swamireddy has joined #openstack-cinder17:24
*** itlinux has joined #openstack-cinder17:25
jungleboyjI still haven't seen Never Ending Story.17:29
*** belmoreira has quit IRC17:29
jungleboyjMy freshman year in College writing course was on A Clockwork Orange.17:29
jungleboyjWas quite the start to college.17:29
mnaserum17:30
mnaserhttps://review.openstack.org/#/c/545201/2/cinder/volume/manager.py L1625 and L1635.. it says its over 80 characters (right at 80), if i indent inwards by 1 space, it says its not properly indented17:30
mnaserdoes anyone mind handholding how clean this up :p17:31
*** luizbag has quit IRC17:32
eharneymnaser: line break after the *, i think17:32
smcginnismnaser: You'll have to do some funky formatting.17:32
mnaseras if it isnt funky enough as is, bleh17:32
smcginnisYeah, what eharney said should work.17:32
mnaserOH17:32
mnaseri think i could refactor it into a simple for loop instead17:32
smcginnisEven better.17:32
mnaserbecause its always <foo>_per_gb which translates to <foo> * <size>17:33
* mnaser wrote that patch late last night17:33
eharneyseems sensible17:33
openstackgerritMohammed Naser proposed openstack/cinder master: Add throughput limits that scale per-GB  https://review.openstack.org/54520117:39
mnaserdone.. should be much cleaner17:40
openstackgerritEric Young proposed openstack/os-brick master: Adding support to extend attached ScaleIO volumes  https://review.openstack.org/54540417:45
*** harlowja has joined #openstack-cinder17:55
*** kaisers_ has quit IRC17:55
*** lbragstad has quit IRC17:55
*** lbragstad has joined #openstack-cinder17:56
*** hemna_ has quit IRC17:57
*** mriedem is now known as mriedem_lunch17:58
*** hemna_ has joined #openstack-cinder18:00
*** mayur_ind has joined #openstack-cinder18:04
*** kaisers_ has joined #openstack-cinder18:04
*** r-daneel has joined #openstack-cinder18:05
*** hemna_ has quit IRC18:09
*** salv-orlando has joined #openstack-cinder18:10
*** hemna_ has joined #openstack-cinder18:11
*** david-lyle has quit IRC18:12
*** dustins has quit IRC18:15
mayur_indhi @smcginnis... where this 'target_conf' option is declared, its not in cinder.conf..... https://review.openstack.org/#/c/539230/3/cinder/volume/drivers/veritas_access/veritas_iscsi.py@4818:16
*** hemna_ has quit IRC18:20
openstackgerritEric Harney proposed openstack/cinder master: QNAP: Pass unit tests with hash randomization enabled  https://review.openstack.org/54538318:23
openstackgerritEric Harney proposed openstack/cinder master: Enable hash randomization in unit tests  https://review.openstack.org/46563618:23
*** AlexeyAbashkin has joined #openstack-cinder18:24
*** hemna_ has joined #openstack-cinder18:25
*** tesseract has quit IRC18:26
openstackgerritMohammed Naser proposed openstack/cinder master: Add throughput limits that scale per-GB  https://review.openstack.org/54520118:27
*** crose has joined #openstack-cinder18:27
*** AlexeyAbashkin has quit IRC18:28
*** r-daneel has quit IRC18:37
*** r-daneel has joined #openstack-cinder18:38
*** psachin has joined #openstack-cinder18:41
*** harlowja has quit IRC18:44
*** harlowja has joined #openstack-cinder18:48
*** kaisers_ has quit IRC18:52
*** r-daneel has quit IRC18:54
*** lpetrut has joined #openstack-cinder18:55
*** r-daneel has joined #openstack-cinder18:58
smcginnismayur_ind: target_port19:08
smcginnismayur_ind: You can search for it here: https://docs.openstack.org/cinder/latest/sample_config.html19:09
*** harlowja has quit IRC19:10
*** hemna_ has quit IRC19:11
*** lpetrut has quit IRC19:12
*** mayur_ind has quit IRC19:12
*** david-lyle has joined #openstack-cinder19:13
*** hemna_ has joined #openstack-cinder19:15
*** lpetrut has joined #openstack-cinder19:20
openstackgerritEric Harney proposed openstack/cinder master: QNAP: Pass unit tests with hash randomization enabled  https://review.openstack.org/54538319:21
openstackgerritEric Harney proposed openstack/cinder master: Enable hash randomization in unit tests  https://review.openstack.org/46563619:22
*** pchavva1 has joined #openstack-cinder19:27
*** dustins has joined #openstack-cinder19:33
*** erlon has quit IRC19:38
openstackgerritMerged openstack/cinder master: Provide a hint when performing a volume action can't find the method  https://review.openstack.org/54537819:40
*** pchavva1 has quit IRC19:42
*** pchavva has quit IRC19:42
openstackgerritEric Harney proposed openstack/cinder master: Tests: Fix python version check  https://review.openstack.org/54544119:44
*** mriedem_lunch is now known as mriedem19:44
mriedemgah devstack doesn't add ssh rules to the default security group, by default19:47
*** harlowja has joined #openstack-cinder19:50
smcginnisHuh, I thought it used to.19:51
mtreinishsmcginnis: I don't think it ever did20:00
mtreinishbut it's been a while since I checked20:00
smcginnisGuess I've always used my real cloud for accessing the guests then.20:01
*** r-daneel_ has joined #openstack-cinder20:02
*** salv-orlando has quit IRC20:03
*** salv-orlando has joined #openstack-cinder20:04
*** r-daneel has quit IRC20:05
*** r-daneel_ is now known as r-daneel20:05
*** dustins_ has joined #openstack-cinder20:08
*** salv-orlando has quit IRC20:08
*** dustins has quit IRC20:08
*** eharney has quit IRC20:13
*** lpetrut has quit IRC20:16
*** lpetrut has joined #openstack-cinder20:17
*** BlackDex has quit IRC20:23
*** salv-orlando has joined #openstack-cinder20:34
Roamer`yep, I think "add a security rule for SSH" was an item in a couple of quick setup guides for OpenStack that I read way back when20:38
*** BlackDex has joined #openstack-cinder20:41
*** andreas_s has joined #openstack-cinder20:43
openstackgerritBrianna Poulos proposed openstack/cinder master: docs: Fix volume encryption link  https://review.openstack.org/54545120:45
*** andreas_s has quit IRC20:47
*** markvoelker_ has quit IRC20:55
*** markvoelker has joined #openstack-cinder20:56
*** liverpooler has quit IRC20:57
*** eharney has joined #openstack-cinder21:05
openstackgerritBrianna Poulos proposed openstack/python-cinderclient master: Update help text for encryption provider  https://review.openstack.org/54545621:08
openstackgerritEric Young proposed openstack/os-brick master: Adding support to extend attached ScaleIO volumes  https://review.openstack.org/54540421:10
*** r-daneel has quit IRC21:10
mriedemyay i found a nova + multiattach bug21:19
ildikovmriedem: should I be happy now? :)21:20
mriedemsure21:20
mriedemit's just a ux problem in nova api21:20
ildikovso what's the bug?21:20
mriedemi was forgetting to specify 2.60 and the very specific error about needing to use 2.60 was catch by a generic Exception block and turned into a much less useful error message21:21
mriedem*caught21:21
mriedemhad to put some debug code in the API to figure it out21:21
*** ganso has quit IRC21:21
mriedemhttps://bugs.launchpad.net/nova/+bug/175006421:23
openstackLaunchpad bug 1750064 in OpenStack Compute (nova) "multiattach volume failures are masked in compute api" [Undecided,New]21:23
ildikovhmm, I thought I remembered seeing that error, but that might've been some trace in tempest21:23
mriedemRoamer`: yeah copied from https://docs.openstack.org/nova/pike/admin/security-groups.html#create-security-group-rules-for-a-cluster-of-instances21:25
smcginnistbarron: Was there expected to be a v2 at some point? https://github.com/openstack/cinder/blob/master/cinder/backup/chunkeddriver.py#L61721:26
tbarronsmcginnis: good question, that was business logic copied from when this was a swift backup only21:27
tbarronmebbe DuncanT ^^^ knows21:27
smcginnistbarron: OK, just thought I would check since I didn't see any signs of anything other than v1. ;)21:27
tbarronsmcginnis: right, at this point I think whatever versioning was imagined is no longer relevant21:28
smcginnistbarron: Given that this code really hasn't been touched since you were working on it, that's probably safe to assume for now. ;)21:29
*** r-daneel has joined #openstack-cinder21:48
*** dustins_ has quit IRC21:55
*** lpetrut has quit IRC22:00
*** edmondsw has quit IRC22:01
*** edmondsw has joined #openstack-cinder22:01
*** psachin has quit IRC22:04
*** edmondsw has quit IRC22:06
*** abishop has quit IRC22:09
*** mchlumsky_ has quit IRC22:18
*** tpsilva has quit IRC22:33
*** zzzeek has quit IRC22:38
*** zzzeek has joined #openstack-cinder22:42
openstackgerritMarga Millet proposed openstack/cinder master: Cinder logs rabbitmq password on connection log  https://review.openstack.org/54548623:11
*** csaikia has quit IRC23:42
*** csaikia has joined #openstack-cinder23:42
openstackgerritMerged openstack/cinder master: Tests: Fix python version check  https://review.openstack.org/54544123:44
*** hemna_ has quit IRC23:50
*** gman-tx has quit IRC23:55
« Thursday, 2018-02-15 Index Saturday, 2018-02-17 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!