Friday, 2017-10-13

*** sbezverk has quit IRC00:06
*** yangyapeng has quit IRC00:11
*** yangyapeng has joined #openstack-cinder00:11
*** yangyapeng has quit IRC00:16
*** sdague has quit IRC00:19
*** lhx_ has joined #openstack-cinder00:24
*** Shunli has joined #openstack-cinder00:42
*** AlexeyAbashkin has joined #openstack-cinder00:42
*** kaisers1 has joined #openstack-cinder00:44
*** kaisers has quit IRC00:46
*** AlexeyAbashkin has quit IRC00:47
*** wanghao_ has quit IRC00:50
*** wanghao has joined #openstack-cinder00:50
*** kwathore_ has joined #openstack-cinder00:57
*** kwathore has quit IRC01:00
*** lhx_ has quit IRC01:06
openstackgerritMerged openstack/cinder master: Deprecate SolidFire Image Caching feature
*** stakeda has joined #openstack-cinder01:10
*** wanghao_ has joined #openstack-cinder01:16
*** Apoorva_ has joined #openstack-cinder01:17
*** wanghao has quit IRC01:18
*** yangyapeng has joined #openstack-cinder01:18
*** yangyapeng has quit IRC01:19
*** yangyapeng has joined #openstack-cinder01:19
*** yangyape_ has joined #openstack-cinder01:20
*** Apoorva has quit IRC01:20
*** Apoorva_ has quit IRC01:21
*** yangyapeng has quit IRC01:24
*** yangyape_ has quit IRC01:28
*** yangyapeng has joined #openstack-cinder01:29
openstackgerritGuy Kim proposed openstack/cinder master: Update configuration names in ITRI DISCO volume driver
*** felipemonteiro has quit IRC01:36
openstackgerritqtlu proposed openstack/cinder master: Fix backup-import error when the deleted record in the same db
*** AlexeyAbashkin has joined #openstack-cinder01:45
*** psachin has joined #openstack-cinder01:46
*** markvoelker has joined #openstack-cinder01:47
*** AlexeyAbashkin has quit IRC01:49
*** dhinesh has quit IRC01:56
openstackgerritPeter Wang proposed openstack/cinder master: Unity: Fix duplicate hosts created with same name
*** chenying_ has joined #openstack-cinder02:06
*** mgagne has quit IRC02:19
*** markvoelker has quit IRC02:21
*** thingee has quit IRC02:22
*** mgagne has joined #openstack-cinder02:22
*** mgagne is now known as Guest5368002:22
*** thingee has joined #openstack-cinder02:23
*** dave-mccowan has quit IRC02:41
openstackgerritGuy Kim proposed openstack/cinder master: Update configuration names in ITRI DISCO volume driver
*** AlexeyAbashkin has joined #openstack-cinder02:43
openstackgerritqtlu proposed openstack/cinder master: Fix backup-import error when the deleted record in the same db
*** AlexeyAbashkin has quit IRC02:48
*** links has joined #openstack-cinder02:51
*** ChubYann has quit IRC02:51
*** nicolasbock has quit IRC03:16
*** markvoelker has joined #openstack-cinder03:18
*** vivsoni has joined #openstack-cinder03:18
*** wes_dillingham has quit IRC03:35
*** edmondsw has joined #openstack-cinder03:38
*** wanghao_ has quit IRC03:40
*** yangyapeng has quit IRC03:40
*** crushil has joined #openstack-cinder03:40
*** wanghao has joined #openstack-cinder03:40
*** yangyapeng has joined #openstack-cinder03:41
*** edmondsw has quit IRC03:45
*** markvoelker has quit IRC03:51
*** Vipul_ has joined #openstack-cinder03:51
*** dhinesh has joined #openstack-cinder04:01
*** mvk has joined #openstack-cinder04:20
*** pcaruana has joined #openstack-cinder04:27
*** mvk has quit IRC04:31
*** gkadam_ has joined #openstack-cinder04:38
*** gkadam_ has quit IRC04:39
*** gkadam has joined #openstack-cinder04:40
openstackgerritMerged openstack/cinder master: Add ability to specify backup driver via class name
*** markvoelker has joined #openstack-cinder04:49
*** armax has quit IRC04:50
*** armax has joined #openstack-cinder04:51
*** armax has joined #openstack-cinder04:51
*** ChubYann has joined #openstack-cinder04:52
*** armax has quit IRC04:52
*** armax has joined #openstack-cinder04:52
*** pcaruana has quit IRC04:56
*** swamireddy has quit IRC05:02
*** cristicalin has joined #openstack-cinder05:10
*** crushil has quit IRC05:17
*** chlong has joined #openstack-cinder05:20
*** markvoelker has quit IRC05:22
*** cristicalin has quit IRC05:23
*** ChubYann has quit IRC05:24
*** gouthamr has quit IRC05:24
*** edmondsw has joined #openstack-cinder05:27
*** pcaruana has joined #openstack-cinder05:30
*** edmondsw has quit IRC05:32
*** swamireddy has joined #openstack-cinder05:37
*** armaan has joined #openstack-cinder05:39
*** pcaruana has quit IRC05:39
*** rcernin has joined #openstack-cinder05:41
*** armaan has quit IRC05:43
*** cristicalin has joined #openstack-cinder05:43
openstackgerritnikesh proposed openstack/cinder master: Kaminario K2: Add non discovery iSCSI multipath
*** spectr has quit IRC05:58
*** spectr has joined #openstack-cinder05:58
*** cah_link__ has joined #openstack-cinder06:09
*** cah_link__ has quit IRC06:10
*** cah_link__ has joined #openstack-cinder06:11
*** cristicalin has quit IRC06:17
*** markvoelker has joined #openstack-cinder06:19
*** lpetrut_ has joined #openstack-cinder06:21
*** e0ne has joined #openstack-cinder06:32
openstackgerritGuy Kim proposed openstack/cinder master: Update configuration names in ITRI DISCO volume driver
*** e0ne has quit IRC06:35
*** e0ne has joined #openstack-cinder06:35
*** wanghao_ has joined #openstack-cinder06:52
*** wanghao has quit IRC06:52
*** markvoelker has quit IRC06:52
*** wanghao has joined #openstack-cinder06:54
*** wanghao_ has quit IRC06:58
*** lhx__ has joined #openstack-cinder06:59
openstackgerritShunei Shiono proposed openstack/cinder master: NEC driver: Never use target info in migration_status
*** pcaruana has joined #openstack-cinder07:07
*** lpetrut_ has quit IRC07:13
openstackgerritMerged openstack/cinder master: [policy in code] Add support for backup resource
*** edmondsw has joined #openstack-cinder07:15
*** gcb has quit IRC07:16
*** armaan has joined #openstack-cinder07:17
*** edmondsw has quit IRC07:20
*** tesseract has joined #openstack-cinder07:22
*** AlexeyAbashkin has joined #openstack-cinder07:24
*** spectr has quit IRC07:26
*** tesseract has quit IRC07:26
*** rcernin has quit IRC07:26
*** pcaruana has quit IRC07:26
*** e0ne has quit IRC07:33
*** mvk has joined #openstack-cinder07:37
*** rcernin has joined #openstack-cinder07:40
*** spectr has joined #openstack-cinder07:40
*** pcaruana has joined #openstack-cinder07:40
*** armaan has quit IRC07:41
*** spectr has quit IRC07:43
*** pcaruana has quit IRC07:44
*** pcaruana has joined #openstack-cinder07:44
*** spectr has joined #openstack-cinder07:44
*** wanghao_ has joined #openstack-cinder07:50
*** wanghao has quit IRC07:50
*** markvoelker has joined #openstack-cinder07:50
*** wanghao has joined #openstack-cinder07:52
tommylikehuhey lbragstad , smcginnis , jungleboyj  could you review this patch?, thanks in advance!07:52
*** wanghao_ has quit IRC07:55
openstackgerritHuang Zhiteng proposed openstack/cinder master: Check available capacity before creating resources
*** AlexeyAbashkin has quit IRC08:10
*** AlexeyAbashkin has joined #openstack-cinder08:11
*** armaan has joined #openstack-cinder08:15
*** gcb has joined #openstack-cinder08:21
*** markvoelker has quit IRC08:23
*** dhinesh has quit IRC08:31
*** itlinux has joined #openstack-cinder08:33
*** itlinux has quit IRC08:36
openstackgerritBoris Pavlovic proposed openstack/cinder master: Switch Rally Task To format V2
*** johnthetubaguy has quit IRC08:39
*** johnthetubaguy has joined #openstack-cinder08:40
*** e0ne has joined #openstack-cinder08:41
*** armaan has quit IRC08:44
*** kaisers1 has quit IRC08:57
*** rcernin has quit IRC08:59
*** spectr has quit IRC08:59
*** pcaruana has quit IRC08:59
*** wanghao has quit IRC09:00
*** wanghao has joined #openstack-cinder09:00
*** AlexeyAbashkin has quit IRC09:03
*** AlexeyAbashkin has joined #openstack-cinder09:03
*** edmondsw has joined #openstack-cinder09:04
openstackgerritGyorgy Szombathelyi proposed openstack/cinder master: Remove deprecated nova config options
*** edmondsw has quit IRC09:08
*** rcernin has joined #openstack-cinder09:12
*** spectr has joined #openstack-cinder09:12
*** pcaruana has joined #openstack-cinder09:13
*** AlexeyAbashkin has quit IRC09:15
*** AlexeyAbashkin has joined #openstack-cinder09:15
*** markvoelker has joined #openstack-cinder09:20
*** Shunli has quit IRC09:22
*** wanghao has quit IRC09:27
*** gcb has quit IRC09:27
*** belmoreira has joined #openstack-cinder09:29
*** dalgaaf has quit IRC09:31
*** chlong has quit IRC09:40
*** vivsoni_ has joined #openstack-cinder09:52
*** vivsoni has quit IRC09:52
*** markvoelker has quit IRC09:53
*** armaan has joined #openstack-cinder09:58
*** itlinux has joined #openstack-cinder10:01
*** yangyapeng has quit IRC10:02
*** sbezverk has joined #openstack-cinder10:14
*** armaan has quit IRC10:17
*** itlinux has quit IRC10:21
*** lpetrut_ has joined #openstack-cinder10:22
*** sdague has joined #openstack-cinder10:24
*** belmoreira has quit IRC10:26
*** psachin has quit IRC10:33
*** pcaruana has quit IRC10:48
*** ociuhandu has quit IRC10:49
*** markvoelker has joined #openstack-cinder10:50
*** nicolasbock has joined #openstack-cinder11:09
*** armaan has joined #openstack-cinder11:10
*** nicolasbock has quit IRC11:14
openstackgerritCiara Stacke proposed openstack/cinder master: VMAX driver - Cannot extend volumes in a replication relationship
*** edmondsw has joined #openstack-cinder11:20
*** ganso has joined #openstack-cinder11:21
*** cah_link__ has quit IRC11:21
openstackgerritCiara Stacke proposed openstack/cinder master: VMAX driver - Cannot extend volumes in a replication relationship
*** cah_link_ has joined #openstack-cinder11:22
*** markvoelker has quit IRC11:23
*** erlon has joined #openstack-cinder11:23
*** alexchadin has joined #openstack-cinder11:24
*** nicolasbock has joined #openstack-cinder11:26
*** armaan has quit IRC11:44
*** armaan has joined #openstack-cinder11:44
*** abishop has joined #openstack-cinder11:56
*** armaan has quit IRC12:01
*** armaan has joined #openstack-cinder12:02
*** tpsilva has joined #openstack-cinder12:02
*** cah_link_ has quit IRC12:03
*** cah_link_ has joined #openstack-cinder12:03
*** mchlumsky has joined #openstack-cinder12:07
*** felipemonteiro has joined #openstack-cinder12:09
*** wes_dillingham has joined #openstack-cinder12:13
*** felipemonteiro has quit IRC12:14
*** markvoelker has joined #openstack-cinder12:15
*** armaan has quit IRC12:21
*** armaan has joined #openstack-cinder12:21
*** eharney has joined #openstack-cinder12:22
*** ociuhandu has joined #openstack-cinder12:30
*** liverpooler has joined #openstack-cinder12:36
*** ociuhandu has quit IRC12:39
*** dave-mccowan has joined #openstack-cinder12:42
*** stakeda has quit IRC12:43
*** links has quit IRC12:45
*** dave-mccowan has quit IRC12:47
*** pcaruana has joined #openstack-cinder12:48
*** alexchadin has quit IRC12:50
*** alexchadin has joined #openstack-cinder12:51
*** dustins has joined #openstack-cinder12:51
*** spectr has quit IRC12:57
*** ociuhandu has joined #openstack-cinder12:58
*** josecastroleon has quit IRC12:59
*** spectr has joined #openstack-cinder13:02
*** mriedem has joined #openstack-cinder13:05
*** jistr is now known as jistr|mtg13:08
*** jmlowe has quit IRC13:12
*** ociuhandu has quit IRC13:14
*** kaisers has joined #openstack-cinder13:20
*** gkadam has quit IRC13:25
*** lhx__ has quit IRC13:26
*** chlong has joined #openstack-cinder13:33
*** armaan has quit IRC13:41
*** armaan has joined #openstack-cinder13:42
*** rcernin has quit IRC13:44
openstackgerritJeremy Zhang proposed openstack/cinder master: Add v3 api-ref for showing all/specific extra specs for a volume type
*** dansmith is now known as superdan13:47
*** zhaochao has quit IRC13:48
*** dave-mccowan has joined #openstack-cinder13:49
*** gouthamr has joined #openstack-cinder13:50
*** gouthamr has quit IRC13:51
*** gouthamr has joined #openstack-cinder13:51
*** dave-mcc_ has joined #openstack-cinder13:55
*** dave-mccowan has quit IRC13:57
*** jmlowe has joined #openstack-cinder13:59
*** mriedem has quit IRC14:03
*** ociuhandu has joined #openstack-cinder14:03
*** mriedem has joined #openstack-cinder14:06
*** Vipul_ has quit IRC14:10
*** armax has joined #openstack-cinder14:11
*** slunkad has quit IRC14:11
*** alexchadin has quit IRC14:12
*** catintheroof has joined #openstack-cinder14:23
*** Dinesh_Bhor has quit IRC14:34
*** dave-mcc_ is now known as dave-mccowan14:35
*** lhx_ has joined #openstack-cinder14:40
*** KeithMnemonic1 has joined #openstack-cinder14:43
*** rcernin has joined #openstack-cinder14:45
*** spectr has quit IRC14:46
*** e0ne has quit IRC14:47
*** KeithMnemonic has quit IRC14:47
*** pooja_jadhav has quit IRC14:51
*** yangyapeng has joined #openstack-cinder14:54
*** jistr|mtg is now known as jistr14:58
*** rcernin has quit IRC15:03
openstackgerritAlan Bishop proposed openstack/cinder-specs master: Migrate ConfKeyManager's fixed-key to Barbican
abishopjungleboyj, smcginnis: ^^ see
smcginnisabishop: Thanks!15:12
*** yangyapeng has quit IRC15:14
*** yangyapeng has joined #openstack-cinder15:15
ildikovjgriffith: mriedem: johnthetubaguy: smcginnis: jungleboyj: stvnoyes: are you around for a quick sync?15:16
*** bhagyashris has quit IRC15:16
jgriffithildikov: 0/15:16
jungleboyjildikov:  :-)15:17
_pewp_jungleboyj (=゚ω゚)ノ15:17
ildikovhi All :)15:18
ildikovlet's dive in15:18
ildikovso good news on new attach patch side, we found a small bug in Nova that caused test failures which is fixed now15:19
openstackgerritJeremy Zhang proposed openstack/cinder master: Add v3 api-ref for showing all/specific extra specs for a volume type
ildikovthere's still a comment in refresh_connection_info from mriedem that was also discussed on the PTG in some extent, need to look into that15:19
*** yangyapeng has quit IRC15:20
johnthetubaguythe change in the ceph key one?15:20
ildikovcomment is on the latest patch set if anyone feels like chiming in15:20
ildikovjohnthetubaguy: ceph key?15:20
*** yangyapeng has joined #openstack-cinder15:21
johnthetubaguyI think that was the use case, something changes in the connection details that needs updating15:21
ildikovjohnthetubaguy: we got into a chat earlier on whether or not refresh_connection_info should result in a volume state change15:22
johnthetubaguyI guess it might involved tearing down that shared connection and bringing it back up again? I am not sure15:22
johnthetubaguy(I guess we should follow the spirit of not hiding thing from each other, just because that's easy)15:23
johnthetubaguyanyways, thats probably a distraction15:23
jgriffithWe had said at one point that connections (attachment connector fields) would be immutable15:23
*** ntpttr_laptop has joined #openstack-cinder15:24
ildikovL122 here for context:
jgriffithif something changes, you need to reset/update etc you get a new attachment instead15:24
jgriffithjohnthetubaguy: I don't know for sure if that relates to what you're saying or not15:24
johnthetubaguykinda does, I could see that making sense15:25
jgriffithbut I guess that ship has kinda sailed at this point15:25
johnthetubaguywhy sailed?15:25
jgriffithwe've sort of broken that model a little already15:25
* jgriffith is checking something....15:25
jgriffithAhhh... nope, we're still holding that contract15:26
jgriffithattachment-update was what I was concerned about but we still enforce that so we're good15:26
* johnthetubaguy phew....15:27
johnthetubaguyno worries, good to check these things15:27
johnthetubaguyit feels quite similar to the migrate volume stuff, only we know its dead to start with15:27
jgriffithThere was a stage in one of the 145 iterations that this came up :)15:27
ildikov145 is such a nice number :)15:28
ildikovMatt's concern on patch set #145 is: This refresh method is called from a few places in the compute manager and is just for getting the latest connection_info from the volume backend in Cinder. This was meant to be a replacement call for os-initialize_connection, which didn't change the volume status. Now it seems we have this new issue to get the latest connection_info.15:30
jungleboyjildikov:  Thanks for referencing my notes.  Nice to know they are used.15:30
ildikovsorry, on #14415:30
ildikovjungleboyj: np, good that we have them :)15:31
johnthetubaguyildikov: seems like we should separate the bug fix into a follow on patch?15:31
johnthetubaguyildikov: i.e. the refetch of the data from cinder, seems fine, as we should stop storing it in Nova in the new flow and never use Nova's cached copy, I presume15:32
ildikovjohnthetubaguy: the bug on reboot is fixed15:32
jgriffithildikov: umm, I'm not sure about that15:32
johnthetubaguyit feels like there are many bug here, and I am thinking about a different one15:32
jgriffithrefectch to Cinder shouldn't cause any status changes, I don't think he's saying that either15:33
johnthetubaguyif adopting the new flow fixes things, that seems all good (I should go read that patch, after this live-migration one)15:33
ildikovone is fixed, if refresh is a concern it's another one15:33
ildikovjgriffith: it uses attachment_update15:33
jgriffithyeah.. don't do that :)15:33
johnthetubaguyoh, so we get the connector, and update that again?15:33
jgriffithThat was what I was saying earlier about "that ship has sailed"15:33
ildikovjgriffith: which in a function called 'refresh' shouldn't be that surprising, but well...15:34
jgriffithif you do that we're defeating one of the last remaining differences in the new Attachment API15:34
jgriffithjohnthetubaguy: yeah, so querying cinder for the stored connector seems fine15:34
jgriffithbut modifying anything IMO is no good15:35
ildikovjgriffith: I think we should be fine here, but thought to point everyone to it, to raise concerns15:35
johnthetubaguyso the bug was where the connector changes, reboot should fix it?15:35
ildikovif any15:35
ildikovjohnthetubaguy: is what got fixed so we're not supposed to fail on reboot anymore15:37
johnthetubaguyOK, I am really confused, but the above patch seems fine15:38
johnthetubaguyignore that for now, going through the live-migrate patch right now15:38
ildikovjohnthetubaguy: ok, sounds good15:38
ildikovjohnthetubaguy: stvnoyes is not around today, but I'm sure he'll fix whatever you find early next week15:39
johnthetubaguyI am kinda hoping I don't find anything, but good to know15:39
ildikovjohnthetubaguy: well, me too, but I didn't want to be too pushy ;)15:40
jgriffithjohnthetubaguy: ditto :)15:40
johnthetubaguyso detach, we have to go into detaching first?15:40
ildikovin the live migrate path you mean?15:40
johnthetubaguyI guess not for an additional attachment15:40
ildikovno, we create the new attachment before deleting the old one AFAIK15:41
johnthetubaguyyep, we certainly do that, I believe15:41
ildikovwe create the attachment in pre_live_migrate15:43
jgriffithildikov: correct (create before delete)15:44
ildikovjgriffith: cool, thanks for confirming15:44
johnthetubaguywhat does volume_api.roll_detaching do?15:44
johnthetubaguyis that rollback the move to the detaching state?15:44
johnthetubaguysorry, I am derailing the catch up here15:46
ildikovyep, it seems to move the volume back to 'in-use'15:48
ildikovfrom 'detaching'15:48
johnthetubaguyso if its already in-use its a no-op or a failure?15:48
mriedemi'm late15:49
mriedemjohnthetubaguy: you're thinking of a different issue15:50
mriedemfor the refresh conn info thing15:50
mriedemthat's bug 145264115:50
openstackbug 1452641 in OpenStack Compute (nova) "Static Ceph mon IP addresses in connection_info can prevent VM startup" [Medium,Confirmed]
johnthetubaguymriedem: yeah, I thought so15:50
johnthetubaguymriedem: thanks15:50
* smcginnis catches up on the pages of scrollback15:50
*** crose has joined #openstack-cinder15:50
ildikovjohnthetubaguy: I would think it's a no-op, but I always have to re-understand that conditional_update function... :/15:50
johnthetubaguyildikov: no worries, its a total edge case, so its just a nit for a follow on patch if thats a problem15:51
ildikovjohnthetubaguy: I like the sound of follow-up :)15:52
ildikovsmcginnis can correct me in the meantime15:52
*** chlong has quit IRC15:53
mriedemso the issue i brought up was,15:53
mriedemnova today calls os-initialize_connection to get the latest connection_info15:53
mriedembut with the new flow, doing that makes cinder expect an attachment_complete call after that, and the volume gets put into attaching status15:53
mriedemthat's not what we (nova) want,15:54
mriedemso we need to know if nova just calls cinder to get the latest version of the attachment and pulls the connection_info off that15:54
mriedemnow that it's stored in cinder15:54
mriedemif that works, it's great15:54
smcginnisildikov: Follow up sounds good to me.15:54
ildikovit should be retrievable from the attachment record15:55
mriedemi'm not sure in what world the connection_info will change 'out of band' on the cinder side15:55
*** AlexeyAbashkin has quit IRC15:55
mriedembut if it does, then cinder is the holder of truth and we're just asking for the latest15:55
ildikovsmcginnis: I mean the question on roll_detaching, sorry :)15:55
johnthetubaguymriedem: +1 just getting it from cinder each time15:55
jgriffithmriedem: yes, and that should work always15:55
jgriffithif there is some case where the connection info can/does change we have a bigger problem.... but would have to just do a new attachment-create to deal with that15:56
johnthetubaguyI think our re-call of initialize_connection breaks some backends, from what I remember15:56
johnthetubaguyjgriffith: see bug 145264115:56
openstackbug 1452641 in OpenStack Compute (nova) "Static Ceph mon IP addresses in connection_info can prevent VM startup" [Medium,Confirmed]
jgriffithI don't know of any case where a backend could/would modify connection info while a volume is supposed to be connected/in-use15:56
johnthetubaguybut that is a separate concern15:56
ildikovDoes this mean we shouldn't call attachment_update in refresh_connection_info?15:57
jgriffithjohnthetubaguy: ummmm... well yeah!15:57
jgriffithYou go in and reconfigure your backend I'm not sure what one would expect15:57
johnthetubaguyshould be attachment_get I think15:57
smcginnisildikov: Making me read now, huh? Yes, I believe you stated correctly about roll_detaching.15:58
ildikovsmcginnis: coolio and sorry for the extra work :)15:58
johnthetubaguyildikov: so you have a link to line in the review this is all about?15:58
ildikovjohnthetubaguy: yeah, that's what I'm reading out of the discussion too15:58
johnthetubaguyin case its in front of you already, I am just being lazy really15:59
ildikovjohnthetubaguy: maybe we could also rename the function to retrieve_connection_info or something...15:59
*** Apoorva has joined #openstack-cinder16:00
*** Apoorva has quit IRC16:01
ildikovjgriffith: any updates on the Cinder side patch with the db migration?16:02
mriedemildikov: yes that's what i'm saying,16:03
mriedemrefresh_connection_info for new style attachment just does an attachment_get()['connection_info']16:03
jgriffithildikov: unit tests all still blowing up16:03
ildikovmriedem: cool, then we're on the same page, thanks16:03
jgriffithtrying to sort it out now, I have fixed the service object up as per mriedem 's suggestions16:03
jgriffithbut still something very bad happening16:04
ildikovone the functional or on the testing side?16:04
ildikovor both?16:04
jgriffithour unit tests aren't unit tests, just incase I haven't mentioned that today :)16:04
jgriffithildikov: functionally it works fine now, unit tests however not so much16:04
ildikovjgriffith: copied and sorry you had to mention that today16:05
ildikovjgriffith: I feel bad now16:05
jgriffithildikov: why?16:05
ildikovjgriffith: it's not a pleasant topic for a Friday neither easily fixable as much as I understand16:06
jgriffithoh LOL16:06
openstackgerritMerged openstack/cinder master: Switch Rally Task To format V2
jgriffithnahh, we've known our unit test design is not so great; nothing new16:06
ildikovdo you think it's fixable by early next week?16:07
ildikovI mean the patch itself :)16:07
jgriffithI'm hoping to fix it today16:07
ildikovsounds great!16:08
ildikovok, I think everyone has homework16:08
ildikovmriedem is off the hook until I fix his comments on the new attach patch :)16:08
johnthetubaguyildikov: mriedem: I added a comment next to yours on that review, does it make any sense?16:09
ildikovjohnthetubaguy: I need to double check when we actually do driver_attach16:11
*** Apoorva has joined #openstack-cinder16:11
ildikovjohnthetubaguy: I think the question here is what design we want as it's doable multiple ways16:12
ildikovjohnthetubaguy: see the comment here:
johnthetubaguyildikov: that seems odd, hmm.16:17
ildikovjohnthetubaguy: that's why I said that it's a design question, like saying that no matter what when we call bdm.attach that means to put the volume to 'in-use'16:18
johnthetubaguyildikov: that might be correct, I think we are re-using some methods in odd ways here, will need more time before I can say anything sensible16:19
ildikovjohnthetubaguy: if we can clean things up that would be great as well16:22
ildikovjohnthetubaguy: we're all in favor of that I would think16:23
*** pcaruana has quit IRC16:23
ildikovjgriffith: mriedem: johnthetubaguy: smcginnis: jungleboyj: ok, I think we're done with the sync for today16:25
smcginnisShip it.16:25
jungleboyjsmcginnis:  ++16:26
ildikovlet's get the homework done early next week so we can clean up the new attach patch and get it landed finally16:26
* ildikov also appreciates all the votes in advance ;)16:26
*** felipemonteiro has joined #openstack-cinder16:27
smcginnisildikov: Homework on the weekend? You're mean.16:31
*** felipemonteiro has quit IRC16:32
ildikovsmcginnis: what's weekend again?16:32
smcginnisildikov: Are you spending yours on an airplane?16:32
ildikovsmcginnis: airplane, workshop, airplane, training, conference, airplane would be my next six days16:33
smcginnisildikov: Fun fun fun. :)16:34
smcginnisAnyone else going to be at All Things Open? eharney?16:34
ildikovildikov: a blast! :)16:34
*** crose has quit IRC16:38
jungleboyjildikov:  Don't know how you do it?16:41
ildikovjungleboyj: don't question it, I have to face the issue here that way :)16:42
*** wes_dillingham has quit IRC16:42
jungleboyjildikov:  Ok, I will leave it to you.  :-)16:43
ildikovjungleboyj: appreciated :)16:46
*** ntpttr_laptop has quit IRC16:48
johnthetubaguyildikov: any chance you could double check my worry on this live-migrate patch to see if I just miss-read the flow?
openstackgerritEric Young proposed openstack/cinder master: ScaleIO Driver: Backup volume via snapshot
ildikovjohnthetubaguy: will do, just saw your comment16:52
*** wes_dillingham has joined #openstack-cinder16:53
johnthetubaguyildikov: thanks16:58
*** ntpttr_laptop has joined #openstack-cinder16:59
*** e0ne has joined #openstack-cinder17:04
ildikovjohnthetubaguy: so in the new attach patch refresh_connection_info calls attachment_update17:07
ildikovjohnthetubaguy: at least currently17:07
ildikovjohnthetubaguy: which is the missing piece in your question there I think17:08
johnthetubaguyildikov: yeah, that would do it17:08
ildikovjohnthetubaguy: however we said we want attachment_get there17:08
johnthetubaguyildikov: well, that wouldn't work in this case of course... sigh17:09
ildikovjohnthetubaguy: yeah, need to check the flow again17:10
johnthetubaguyildikov: so in this case we have done attachment_create, about to do driver_attach, so we need the middle bit17:10
ildikovjohnthetubaguy: if we call bdm.attach we have an attachment_update call there17:10
*** jgriffith has quit IRC17:11
johnthetubaguyildikov: but the driver attach is part of the pre_live_migration call right now17:12
*** John has joined #openstack-cinder17:12
*** John is now known as jgriffith17:12
jungleboyjildikov: jgriffith patrickeast  Good job covering for me in the RedHat PTG Interview.  :-)17:13
jungleboyjThank you.  It turned out very well.17:13
ildikovjohnthetubaguy: this one you mean: driver.pre_live_migration?17:13
johnthetubaguyildikov: yes, calls _connect_volume17:14
ildikovjungleboyj: that video is hilarious :D17:14
ildikovjohnthetubaguy: ok, got you now17:14
*** masber has quit IRC17:14
*** mvk has quit IRC17:14
johnthetubaguyildikov: I added a suggested fix on the patch17:16
ildikovjohnthetubaguy: cool, I will check17:17
jungleboyjildikov:  Yep, but it got the job done.    I could hear your phone ding in the background I think as I was trying to tell you I was coming.17:17
ildikovjohnthetubaguy: doesn't sound to be a complicated problem17:17
johnthetubaguyildikov: in isolation its a simple fix I think17:17
ildikovjohnthetubaguy: I will try whether tests are failing without that fix17:18
ildikovjohnthetubaguy: if they do we should get the live_migrate patch merged and fix it in the new attach patch17:18
ildikovjohnthetubaguy: testing and follow-up is easier that way17:18
johnthetubaguyildikov: that is the odd bit, how do the tests pass, confuses me17:18
johnthetubaguyoh wait, they pass as the follow on patch fixes it, I get it now17:18
ildikovjungleboyj: I think my phone was silent17:19
jungleboyjHmmm, ok.  Someone's phone dinged.17:19
ildikovjohnthetubaguy: the new attach is not enabled in the live_migrate patch, so it runs the old flow17:19
ildikovjohnthetubaguy: the new flow patch depends on the live_migrate one, so it passes the tests as it has everything17:19
*** lpetrut_ has quit IRC17:19
johnthetubaguyildikov: they did tests by adding the two patches together I think17:21
ildikovjohnthetubaguy: yes17:21
johnthetubaguyyeah, we just said the same thing differently, doh17:21
ildikovjohnthetubaguy: so if this part is the issue we should be able to catch it with tests I would think17:21
johnthetubaguyOK, so I think I have this all in my head, must be time for the weekend (face palm)17:22
*** dhinesh has joined #openstack-cinder17:22
ildikovjohnthetubaguy: need to go and find my flight, but I will look into a solution17:22
ildikovjohnthetubaguy: true :)17:22
johnthetubaguy++ have a safe flight17:22
*** lhx_ has quit IRC17:22
ildikovjohnthetubaguy: thanks!17:22
openstackgerritMerged openstack/python-brick-cinderclient-ext master: Updated from global requirements
johnthetubaguyildikov: I added comments saying possible just pull out that change in the other patch, and depend on that17:22
ildikovjohnthetubaguy: I will check, but would rather not overcomplicate it for review sanity perspective17:23
johnthetubaguyildikov: I am happy to go with what mriedem feels, my problem is clearly fixed in the follow up patch17:24
ildikovjohnthetubaguy: +117:24
ildikovjohnthetubaguy: thanks for the review17:25
*** catintheroof has quit IRC17:25
*** catintheroof has joined #openstack-cinder17:25
*** dhinesh_ has joined #openstack-cinder17:25
johnthetubaguyildikov: apologies on the wait, had a few things in the way this week17:26
*** AlexeyAbashkin has joined #openstack-cinder17:27
*** mvk has joined #openstack-cinder17:27
*** AlexeyAbashkin has quit IRC17:28
*** dhinesh has quit IRC17:29
*** catintheroof has quit IRC17:29
ildikovjohnthetubaguy: it's ok, I know how that goes17:35
*** e0ne has quit IRC17:37
*** armax has quit IRC17:42
*** nicolasbock has quit IRC17:42
*** sgu has joined #openstack-cinder17:44
openstackgerritJohn Griffith proposed openstack/cinder master: Fix migration 112 to use live_data_migration API
*** dosaboy has quit IRC17:49
*** armax has joined #openstack-cinder17:52
*** lpetrut_ has joined #openstack-cinder17:55
jgriffithservice queries are still failing on that ^^17:56
jgriffithas is an SQLite DateTime entry17:57
jgriffithI've screwed something up pretty good :)17:57
ildikovjgriffith: good job! :)18:10
*** armaan has quit IRC18:13
*** Guest53680 is now known as mgagne18:14
*** mgagne has quit IRC18:14
*** mgagne has joined #openstack-cinder18:14
*** eharney has quit IRC18:20
*** ntpttr_laptop has quit IRC18:22
openstackgerritMerged openstack/cinder master: Update configuration names in ITRI DISCO volume driver
openstackgerritMerged openstack/cinder master: Updated from global requirements
jgriffithildikov: haha!  Thanks!!  I've outdone myself :)18:25
*** mriedem1 has joined #openstack-cinder18:29
*** mriedem has quit IRC18:31
*** eharney has joined #openstack-cinder18:31
*** mriedem1 is now known as mriedem18:32
*** eharney has quit IRC18:32
*** eharney has joined #openstack-cinder18:32
*** jmlowe has quit IRC18:33
*** catintheroof has joined #openstack-cinder18:33
*** armax_ has joined #openstack-cinder18:40
*** cah_link_ has quit IRC18:40
*** armax has quit IRC18:42
*** armax_ is now known as armax18:42
*** gouthamr has quit IRC18:47
*** dosaboy has joined #openstack-cinder18:48
*** felipemonteiro has joined #openstack-cinder18:48
*** dosaboy has quit IRC18:48
*** dosaboy has joined #openstack-cinder18:50
*** felipemonteiro_ has joined #openstack-cinder18:50
*** felipemonteiro has quit IRC18:53
*** ChubYann has joined #openstack-cinder19:10
*** lpetrut_ has quit IRC19:15
*** armaan has joined #openstack-cinder19:17
*** sgu has quit IRC19:19
*** gouthamr has joined #openstack-cinder19:22
*** gouthamr_ has joined #openstack-cinder19:24
*** gouthamr has quit IRC19:26
ildikovjgriffith: Friday the 13th and the next workday is Monday19:39
*** ntpttr_laptop has joined #openstack-cinder19:39
*** AlexeyAbashkin has joined #openstack-cinder19:40
ildikovjgriffith: so you didn't really have a chance :)19:40
ildikovjgriffith: by Tuesday it should all work though, magically :)19:40
jgriffithildikov: not holding my breath on that, but sounds nice :)19:44
*** AlexeyAbashkin has quit IRC19:44
ildikovjgriffith: no guarantees, I'm still practicing this predicting the future thing... :)19:49
*** wes_dillingham has quit IRC19:53
openstackgerritMerged openstack/cinder master: Add v3 api-ref for showing all/specific extra specs for a volume type
*** jmlowe has joined #openstack-cinder20:02
*** felipemonteiro__ has joined #openstack-cinder20:06
*** felipemonteiro_ has quit IRC20:09
*** masber has joined #openstack-cinder20:10
*** sdague has quit IRC20:13
*** masber has quit IRC20:15
jgriffithahh... good news is I now know precisely what's blowing up20:17
*** shaner has quit IRC20:18
*** KeithMnemonic1 has quit IRC20:19
*** shaner has joined #openstack-cinder20:21
*** Apoorva_ has joined #openstack-cinder20:30
smcginnisYour code?20:31
*** Apoorva has quit IRC20:34
ildikovsmcginnis: be nice!20:36
smcginnisGotta give jgriffith crap when I can!20:37
ildikovsmcginnis: and say these things *after* he uploaded the fixed version ;)20:37
smcginnisildikov: Oh OK, save the heckling until later. Got it. ;)20:37
jgriffithsmcginnis: well... yeah :)20:37
smcginnisIt's just not as fun though!20:38
jgriffithmeh, I'll just abandon the patch and lete somebody else deal with it :)20:38
ildikovsmcginnis: sorry for killing the joy20:38
smcginnisjgriffith: You're going to cause ildikov to travel half way around the world to stab me to death now.20:38
jgriffithsmcginnis: Perhaps I'll upgrade her seat for her :)20:39
ildikovjgriffith: YES PLEASE! lol :)20:39
jgriffithWell, I would but then I could be considered an accessory to murder, and well; that's no good20:40
jgriffithbesides smcginnis is an alright guy20:40
jgriffithnow... if you want to just mame him a bit... then we can talk :)20:40
smcginnisSheesh, guess I'm going to have to be looking over my shoulder in Sydney now!20:41
jgriffithsleep with one eye open mcginnis!!!20:44
smcginnisConsidering how poisonous nearly everything is in Australia, I think I'll just stay awake in this case.20:45
jgriffithsmcginnis: yeah, that's my plan as well!!!20:45
*** edmondsw has quit IRC20:45
jgriffithgoing to put myself in a giant zip lock bag with some air holes20:46
*** Apoorva_ has quit IRC20:46
jgriffithon a *somewhat* more serious note....20:46
jgriffithThe problem is that patch adds a on the service_from_db_object call20:47
jgriffithturns out that a huge number of the tests hit that20:47
jgriffithand there's nothing mocking or patching it20:48
jgriffithso the makes a call to the db and blows up20:48
smcginnisI'm sure we have many, many more cases where tests end up actually doing things they shouldn't that we don't even know about.20:48
jgriffithsmcginnis: yeah, we kinda test a LOT of stuff in the unit tests without even knowing it20:49
jgriffithI need to apply this:  self.patch('', return_value=None)20:51
ildikovsmcginnis: jgriffith: I was just going for the upgrade20:51
jgriffithto every single test :)20:51
ildikovsmcginnis: jgriffith: not a killer type, or not in that meaning at least :)20:51
jgriffithildikov: I don't think anybody thought you were really capable of such things20:52
ildikovjgriffith: you can use it on class level20:52
ildikovjgriffith: the mock I mean20:52
jgriffithildikov: sure20:52
jgriffithildikov: 255 failing tests...20:52
jgriffithadded it to volume class, get's me down around 150 or so20:52
ildikovjgriffith: with using it on class level?20:52
jgriffithI guess I could just pick through all fo them20:53
jgriffithildikov: yes, unless you mean higher level class20:53
jgriffithbut then I'm worried I'm masking somethign that I shouldn't be20:53
ildikovjgriffith: by mocking 'save'?20:54
ildikovjgriffith: the unit test shouldn't access the DB on the first place20:54
jgriffithyou know.. the whole "unintended consequences" thing20:54
jgriffithdid you seriously just say the unit tests shouldn't access the DB?20:55
ildikovyou're scared of accidentally fixing the Cinder tests now? :)20:55
ildikovI know, I'm such a next-level genius today :)20:56
jgriffithIt's a swarm of magic code and I don't want to be the last person to touch it when it all blows up :)20:56
smcginnisBack to our definitions of unit tests again.20:56
ildikovif mocking out 'save' blows up Cinder then we might as well look for another profession20:57
ildikovlike making comedy on project status reports in front of a camera :)20:57
jgriffithhaha... my answer to Zane would be "people that need a place to run an openstack CI" :)20:58
*** abishop has quit IRC20:58
jgriffithildikov: that's not a bad idea, I'll take your advice and do just that20:59
smcginnisMight be interesting to see what else blows up. :)20:59
ildikovjgriffith: feel free to blame me if it does not work out21:00
jgriffithall the things that do a fail21:00
ildikovwhat are they failing with exactly?21:00
jgriffithildikov: if you really want to know you should pull down my patch and try it for yourself :)21:01
jgriffithsee what I did there... :)21:01
*** mchlumsky has quit IRC21:02
*** catintheroof has quit IRC21:10
*** wes_dillingham has joined #openstack-cinder21:13
*** dhinesh_ has quit IRC21:25
*** armax has quit IRC21:25
ildikovjgriffith: :P :)21:26
*** dustins has quit IRC21:28
mriedemjungleboyj: jgriffith: smcginnis: cburgess: took me about 7 months but here is the spec
smcginnismriedem: Nice. Seems like the requests for that have been increasing.21:31
mriedemthere is a todo in there that has to be sorted out21:31
smcginnisOr there's a small number that are getting more vocal.21:31
mriedemi'm going to send to the ops and dev list for discussion and awareness21:31
mriedemand then i'm off21:31
*** wes_dillingham has quit IRC21:32
*** dave-mccowan has quit IRC21:34
*** felipemonteiro__ has quit IRC21:36
*** armax has joined #openstack-cinder21:36
*** AlexeyAbashkin has joined #openstack-cinder21:39
*** Apoorva has joined #openstack-cinder21:41
*** wes_dillingham has joined #openstack-cinder21:44
*** AlexeyAbashkin has quit IRC21:44
*** wes_dillingham has quit IRC22:09
*** tpsilva has quit IRC22:10
*** gouthamr_ has quit IRC22:23
*** Swimingly has quit IRC22:25
*** Swimingly has joined #openstack-cinder22:26
*** AlexeyAbashkin has joined #openstack-cinder22:39
*** AlexeyAbashkin has quit IRC22:43
*** Apoorva has quit IRC22:53
*** Apoorva has joined #openstack-cinder22:53
*** wes_dillingham has joined #openstack-cinder23:11
*** ntpttr_laptop has quit IRC23:12
*** wes_dillingham has quit IRC23:13
*** gouthamr has joined #openstack-cinder23:30
*** AlexeyAbashkin has joined #openstack-cinder23:39
*** gouthamr_ has joined #openstack-cinder23:43
*** AlexeyAbashkin has quit IRC23:43
*** gouthamr has quit IRC23:46
*** Apoorva has quit IRC23:46
*** superdan is now known as dansmith23:47
*** markvoelker has quit IRC23:49
*** gouthamr_ has quit IRC23:57

Generated by 2.15.3 by Marius Gedminas - find it at!