Monday, 2016-08-15

*** ducttape_ has quit IRC00:00
*** yangyapeng has quit IRC00:02
*** raunak has quit IRC00:03
*** raunak has joined #openstack-cinder00:05
*** jiaohaolin has quit IRC00:07
*** alonma has joined #openstack-cinder00:17
*** alonma has quit IRC00:22
*** jwcroppe has joined #openstack-cinder00:23
*** jwcroppe has quit IRC00:28
*** raunak has quit IRC00:37
*** mriedem has quit IRC00:58
*** cheneydc has joined #openstack-cinder00:59
*** ducttape_ has joined #openstack-cinder01:01
*** bswartz has joined #openstack-cinder01:02
*** cheneydc has quit IRC01:05
*** ducttape_ has quit IRC01:05
*** lixiaoy1 has quit IRC01:10
*** lixiaoy1 has joined #openstack-cinder01:11
lixiaoy1jgriffith: Hi John, may I talk to you about your concerns about these two patches: and ?01:12
*** bswartz has quit IRC01:13
*** TaylorHuang has joined #openstack-cinder01:14
*** TaylorH has quit IRC01:17
*** yangyapeng has joined #openstack-cinder01:17
openstackgerritPatrick East proposed openstack/cinder: Add tempest tests for Consistency Groups
*** EinstCrazy has joined #openstack-cinder01:23
*** June has quit IRC01:30
*** laughterwym has joined #openstack-cinder01:32
*** Lee1092 has joined #openstack-cinder01:35
*** laughterwym has quit IRC01:40
*** laughterwym has joined #openstack-cinder01:40
*** laughterwym has quit IRC01:45
*** cknight has left #openstack-cinder01:46
*** ducttape_ has joined #openstack-cinder01:46
*** ducttape_ has quit IRC01:56
openstackgerrityangweiwei proposed openstack/cinder: Update db operation of cinder in "extra_specs"
*** jamielennox is now known as jamielennox|away02:10
*** yangyape_ has joined #openstack-cinder02:11
*** yangyapeng has quit IRC02:14
*** yangyape_ has quit IRC02:23
*** yangyapeng has joined #openstack-cinder02:24
*** yangyape_ has joined #openstack-cinder02:24
*** jwcroppe has joined #openstack-cinder02:26
*** yangyapeng has quit IRC02:29
*** jamielennox|away is now known as jamielennox02:30
*** jwcroppe has quit IRC02:31
*** ducttape_ has joined #openstack-cinder02:32
*** lixiaoy1 has quit IRC02:40
*** lickitysplitted has left #openstack-cinder02:42
openstackgerritxiexs proposed openstack/cinder: Refactor the usage of save_and_reraise_exception
*** ducttape_ has quit IRC03:02
*** gcb has joined #openstack-cinder03:02
openstackgerritArsen Chen proposed openstack/cinder: Revise Synology DSM storage driver
*** laughterwym has joined #openstack-cinder03:31
*** laughterwym has quit IRC03:36
*** laughterwym has joined #openstack-cinder03:37
*** laughterwym has quit IRC03:41
*** Lee1092 has quit IRC03:43
*** laughterwym has joined #openstack-cinder03:47
*** laughterwym has quit IRC03:50
*** laughterwym has joined #openstack-cinder03:50
*** EinstCrazy has quit IRC03:53
*** laughterwym has quit IRC03:55
*** ducttape_ has joined #openstack-cinder04:02
*** Asaithambi has joined #openstack-cinder04:07
*** ducttape_ has quit IRC04:07
*** alonma has joined #openstack-cinder04:19
*** chlong has joined #openstack-cinder04:21
*** alonma has quit IRC04:23
*** jwcroppe has joined #openstack-cinder04:29
*** alonma has joined #openstack-cinder04:31
*** gouthamr has quit IRC04:33
*** EinstCrazy has joined #openstack-cinder04:33
*** jwcroppe has quit IRC04:33
*** alonma has quit IRC04:36
*** alonma has joined #openstack-cinder04:38
*** alonma has quit IRC04:43
openstackgerritxiexs proposed openstack/cinder: Remove the using of dict.get() in assertIsNone
*** Asaithambi has quit IRC04:44
*** alonma has joined #openstack-cinder04:46
*** laughterwym has joined #openstack-cinder04:47
*** laughterwym has quit IRC04:50
*** alonma has quit IRC04:50
*** laughterwym has joined #openstack-cinder04:51
*** alonma has joined #openstack-cinder04:53
*** EinstCrazy has quit IRC04:54
*** EinstCrazy has joined #openstack-cinder04:55
*** laughterwym has quit IRC04:55
*** alonma has quit IRC04:57
*** EinstCrazy has quit IRC04:59
*** EinstCrazy has joined #openstack-cinder05:00
*** ducttape_ has joined #openstack-cinder05:04
*** alonma has joined #openstack-cinder05:06
*** bswartz has joined #openstack-cinder05:07
*** ducttape_ has quit IRC05:08
*** lkuchlan has joined #openstack-cinder05:09
*** alonma has quit IRC05:11
*** pbandark has joined #openstack-cinder05:14
*** crose has quit IRC05:15
*** crose has joined #openstack-cinder05:18
*** GB21 has joined #openstack-cinder05:21
*** rcernin has joined #openstack-cinder05:25
*** EinstCrazy has quit IRC05:26
*** EinstCrazy has joined #openstack-cinder05:27
*** EinstCra_ has joined #openstack-cinder05:30
*** EinstCrazy has quit IRC05:31
*** lixiaoy1 has joined #openstack-cinder05:38
*** ebalduf has quit IRC05:44
*** rkrum has joined #openstack-cinder05:47
*** liverpooler has joined #openstack-cinder05:59
*** lkuchlan has quit IRC06:02
*** ducttape_ has joined #openstack-cinder06:04
*** alonma has joined #openstack-cinder06:06
*** ducttape_ has quit IRC06:09
*** alonma has quit IRC06:10
*** jmccrory is now known as jmccrory_away06:12
*** alonma has joined #openstack-cinder06:13
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Imported Translations from Zanata
*** alonma has quit IRC06:18
*** alonma has joined #openstack-cinder06:20
*** jwcroppe has joined #openstack-cinder06:23
*** alonma has quit IRC06:24
*** alonma has joined #openstack-cinder06:27
*** jwcroppe has quit IRC06:28
*** lixiaoy1 has quit IRC06:29
*** lixiaoy1 has joined #openstack-cinder06:30
*** alonma has quit IRC06:31
openstackgerritxiexs proposed openstack/cinder: Refactor the using of dict.get() in the test assertion
*** alonma has joined #openstack-cinder06:35
*** david-lyle has joined #openstack-cinder06:36
*** pcaruana has joined #openstack-cinder06:36
*** houming has joined #openstack-cinder06:38
*** david-lyle_ has quit IRC06:39
*** alonma has quit IRC06:39
*** alonma has joined #openstack-cinder06:42
*** alonma has quit IRC06:46
*** belmoreira has joined #openstack-cinder06:48
*** alonma has joined #openstack-cinder06:49
*** alonma has quit IRC06:53
*** gcb has quit IRC06:56
*** alonma has joined #openstack-cinder06:56
*** alonma has quit IRC07:00
*** Thelo has joined #openstack-cinder07:02
*** lkuchlan has joined #openstack-cinder07:03
*** alonma has joined #openstack-cinder07:03
*** ducttape_ has joined #openstack-cinder07:05
*** rkrum has quit IRC07:07
*** alonma has quit IRC07:07
*** alonma has joined #openstack-cinder07:08
*** chlong has quit IRC07:08
*** ducttape_ has quit IRC07:09
*** Thelo has quit IRC07:10
*** Thelo has joined #openstack-cinder07:12
openstackgerritMatan Sabag proposed openstack/cinder: Sending ScaleIO volume id in attach and detach volume
*** arecknag has joined #openstack-cinder07:16
*** Thelo has left #openstack-cinder07:17
*** savihou has joined #openstack-cinder07:21
*** chlong has joined #openstack-cinder07:25
*** lpetrut has joined #openstack-cinder07:34
*** GB21 has quit IRC07:38
*** Lee1092 has joined #openstack-cinder07:40
*** v12aml has joined #openstack-cinder07:41
*** alonma has quit IRC07:45
*** alonma has joined #openstack-cinder07:50
*** GB21 has joined #openstack-cinder07:53
*** yangyape_ has quit IRC07:54
*** yangyapeng has joined #openstack-cinder07:57
*** zzzeek has quit IRC08:00
*** lixiaoy1 has quit IRC08:01
*** sdake has joined #openstack-cinder08:01
*** zzzeek has joined #openstack-cinder08:02
*** savihou has quit IRC08:04
*** yangyapeng has quit IRC08:06
*** yangyapeng has joined #openstack-cinder08:06
*** ducttape_ has joined #openstack-cinder08:06
*** chlong has quit IRC08:07
openstackgerritDavanum Srinivas (dims) proposed openstack/cinder: [WIP] Testing latest u-c
*** ducttape_ has quit IRC08:11
*** e0ne has joined #openstack-cinder08:15
*** sdake has quit IRC08:15
*** alonma has quit IRC08:21
*** alonma has joined #openstack-cinder08:22
*** houming has quit IRC08:30
*** Thelo has joined #openstack-cinder08:34
*** e0ne has quit IRC08:35
*** sdake has joined #openstack-cinder08:35
*** e0ne has joined #openstack-cinder08:37
*** houming has joined #openstack-cinder08:40
*** markvoelker has joined #openstack-cinder08:41
*** lixiaoy1 has joined #openstack-cinder08:41
*** sdake has quit IRC08:43
*** markvoelker has quit IRC08:45
*** lpetrut has quit IRC08:47
*** laughterwym has joined #openstack-cinder08:50
*** houming has quit IRC08:50
*** alonma has quit IRC08:51
*** alonma has joined #openstack-cinder08:51
*** Thelo has quit IRC08:52
*** mvk has joined #openstack-cinder08:56
*** houming has joined #openstack-cinder08:57
*** lpetrut has joined #openstack-cinder08:58
guyDuncanT: Hi DuncanT, are you here?09:00
*** ducttape_ has joined #openstack-cinder09:07
*** ducttape_ has quit IRC09:11
*** lixiaoy1 has quit IRC09:12
*** lixiaoy1 has joined #openstack-cinder09:12
nikeshmDuncanT:  hi, can you give +2 on
*** avishay has joined #openstack-cinder09:19
DuncanTguy: Yes, briefly at least09:20
DuncanTnikeshm: I'll take a look09:20
avishayHi all, in cinder/brick/local_dev/, we use self._execute() for everything except in method get_all_volume_groups where we use putils.execute(). Is this on purpose?09:20
guyDuncanT: ok. I took a look at chunked driver09:20
DuncanTnikeshm: What CI covers the K2 array?09:21
DuncanTguy: Great. Does it look useful?09:21
guyDuncanT: I think that we cannot use it for disco storage system. Because it needs to create a container to store the metadata file inside. By the way,  Ceph doesn't use the chunked driver. Google does.09:21
DuncanTHmmm, I'll look at updating the ceph driver then, thanks for the heads up09:22
nikeshmDuncanT: run-Kaminario K2 CI09:22
DuncanTguy: 'container' in this case means 'anywhere you can storage a metadata file' which isn't necessarily a container in the object store sense09:23
DuncanTnikeshm: Ok, it passed. Thanks09:24
DuncanTnikeshm: In future, mentioning the full array name, rather than just 'K2' int he commit message would generally help those of use with 80 drivers to keep track of09:24
*** savihou has joined #openstack-cinder09:24
nikeshmDuncanT: sure09:25
DuncanTnikeshm: This one is +2 though09:25
DuncanTI've got to step out for lunch, will see messages on my return09:28
guyDuncanT: ah ok, I thought it was the container in the object store09:28
guyDuncanT: Thank you DuncanT09:28
*** lpetrut has quit IRC09:30
DuncanTguy: The method names are that way because it was written for an object store, but really, it just needs somewhere to store the json blob where it can be got back. You could even write it to the end of the volume and put the offset in a binary header, I guess...09:30
guyDuncanT: Ok thanks I will think about it.09:32
*** lixiaoy1 has quit IRC09:39
*** lixiaoy1_ has joined #openstack-cinder09:39
*** asselin has joined #openstack-cinder09:45
*** sdake has joined #openstack-cinder09:46
*** asselin_ has quit IRC09:48
*** Asaithambi has joined #openstack-cinder09:50
*** alonma has quit IRC09:51
*** alonma has joined #openstack-cinder09:51
*** laughterwym has quit IRC09:55
*** alonma has quit IRC09:56
*** avishay has quit IRC09:58
*** alonma has joined #openstack-cinder09:59
*** Lallana has quit IRC10:01
*** alonma has quit IRC10:03
*** alonma has joined #openstack-cinder10:06
*** ducttape_ has joined #openstack-cinder10:08
*** alonma has quit IRC10:10
*** ducttape_ has quit IRC10:12
*** alonma has joined #openstack-cinder10:13
*** avishay has joined #openstack-cinder10:14
*** alonma has quit IRC10:14
*** alonma has joined #openstack-cinder10:14
*** EinstCra_ has quit IRC10:20
*** jwcroppe has joined #openstack-cinder10:20
*** yangyapeng has quit IRC10:20
*** jwcroppe has quit IRC10:27
*** sdague has joined #openstack-cinder10:31
*** alonma has quit IRC10:40
*** alonma has joined #openstack-cinder10:41
*** markvoelker has joined #openstack-cinder10:42
*** alonma has quit IRC10:44
*** alonma has joined #openstack-cinder10:44
*** lpetrut has joined #openstack-cinder10:45
*** markvoelker has quit IRC10:47
*** lpetrut has quit IRC11:06
*** ducttape_ has joined #openstack-cinder11:08
jgriffithavishay: yes, not there actually a few uses in there and they are all static methods11:09
jgriffithavishay: that's to say "yes, it's inentional" and that it is because those are static methods (get_all_volume_groups, get_lvm_version, get_lv_info and get all physcial_volumes)11:11
jgriffithlixiaoy1_: pong11:11
*** alonma has quit IRC11:12
*** ducttape_ has quit IRC11:13
avishayjgriffith: gotcha, ok thank you11:13
*** alonma has joined #openstack-cinder11:13
*** yangyapeng has joined #openstack-cinder11:23
*** xiexs has joined #openstack-cinder11:42
*** gcb has joined #openstack-cinder11:45
nikeshmDuncanT:  jenkins gave -2 on
nikeshmshall i do reverify11:49
*** sdake has quit IRC12:02
xiexssounds now it's just recheck12:02
jgriffithxiexs: you're correct12:03
jgriffiththere's no distinction any longer12:03
*** edmondsw has joined #openstack-cinder12:04
openstackgerritWenjun Wang proposed openstack/cinder-specs: Make quota-class working better
*** sdake has joined #openstack-cinder12:06
*** JoseMello has joined #openstack-cinder12:09
*** ducttape_ has joined #openstack-cinder12:09
*** gouthamr has joined #openstack-cinder12:10
*** gcb has quit IRC12:12
*** ducttape_ has quit IRC12:12
*** gcb has joined #openstack-cinder12:13
*** ducttape_ has joined #openstack-cinder12:13
*** jwcroppe has joined #openstack-cinder12:15
xiexsHi, all, do we have a plan for the "centralize-config-options"? Just like nova does.12:15
nikeshmxiexs jgriffith : ok thanks12:16
xiexsI submitted a bp for it, please give me some advice if you have time. Thanks in advance.12:16
xiexsThe URL is
xiexsnikeshm: you're welcome:)12:17
*** avishay has quit IRC12:19
*** jwcroppe has quit IRC12:26
*** markvoelker has joined #openstack-cinder12:30
*** GB21 has quit IRC12:35
*** mtanino has joined #openstack-cinder12:35
*** ducttape_ has quit IRC12:35
*** sdake has quit IRC12:38
*** alonma has quit IRC12:41
*** alonma has joined #openstack-cinder12:42
*** alonma has quit IRC12:43
*** alonma has joined #openstack-cinder12:43
*** julim has joined #openstack-cinder12:44
*** rkrum has joined #openstack-cinder12:47
*** rlrossit has joined #openstack-cinder12:49
*** laughterwym has joined #openstack-cinder12:50
*** pauloewerton has joined #openstack-cinder12:50
*** amoralej|off has quit IRC12:50
*** amoralej has joined #openstack-cinder12:53
*** avishay has joined #openstack-cinder12:54
*** gcb has quit IRC12:55
DuncanTxiexs: Funny you should mention that, I was just reading cinder/ and now want to sit and gibber for a while. We've definitely lost the plot entirely on that one12:55
*** xyang1 has joined #openstack-cinder12:56
*** alonma has quit IRC13:00
*** julim has quit IRC13:01
*** alonma has joined #openstack-cinder13:03
*** alonma has joined #openstack-cinder13:03
*** houming has quit IRC13:04
*** julim has joined #openstack-cinder13:05
*** avishay has quit IRC13:05
*** rkrum has quit IRC13:10
*** rkrum has joined #openstack-cinder13:11
*** ducttape_ has joined #openstack-cinder13:14
*** rkrum has quit IRC13:16
*** jwcroppe has joined #openstack-cinder13:19
*** lkuchlan has quit IRC13:19
openstackgerritxing-yang proposed openstack/cinder: Differentiate thick and thin provisioning
*** dustins has joined #openstack-cinder13:23
*** GB21 has joined #openstack-cinder13:24
openstackgerritMerged openstack/cinder: Make the volume-manage doc be more general
*** porrua has joined #openstack-cinder13:27
*** belmoreira has quit IRC13:31
*** ducttape_ has quit IRC13:33
*** mriedem has joined #openstack-cinder13:34
*** akerr has joined #openstack-cinder13:37
*** jwcroppe has quit IRC13:37
*** Asaithambi has quit IRC13:37
*** GB21 has quit IRC13:41
*** rlrossit has quit IRC13:48
*** belmoreira has joined #openstack-cinder13:50
*** rlrossit_ has joined #openstack-cinder13:52
*** chlong has joined #openstack-cinder13:52
*** diablo_rojo has joined #openstack-cinder13:56
ildikovjgriffith: hi13:57
jgriffithildikov: morning13:57
jgriffithildikov: or evening :)13:57
ildikovjgriffith: afternoon :)13:57
ildikovjgriffith: there's a meeting popped up in parallel with hours13:57
jgriffithIt was bound to be one of those13:57
* ildikov needs more coffee...13:58
*** dims has quit IRC13:59
ildikovhaha, no worries I usually talk to people, who start with good morning in my afternoon time, so I tend to believe by now that it's still morning :)13:59
jgriffithildikov: ha14:00
jgriffithildikov: so do you want to reschedule?14:00
ildikovjgriffith: I can start our meeting today and try to follow, I was wondering of you could facilitate?14:00
jgriffithI might be able to do that14:01
ildikovif the time slot is not comfortable for you either we can try a reschedule14:01
*** Asaithambi has joined #openstack-cinder14:01
ildikovit's just never easy to find time14:01
jgriffithvery true14:01
ildikovthat fits for everyone... :S14:01
jgriffithwhat time are we set for now again?14:02
jgriffithOk, sure14:02
ildikovBTW, I'll be at OpenStackEast next week14:03
smcginnisildikov: Will you also be at the operators midcycle?14:04
ildikovI saw you on the schedule, so I hoped we can catch up14:04
ildikovsmcginnis: I registered, so that's the plan, yes14:04
*** ducttape_ has joined #openstack-cinder14:04
smcginnisildikov: Cool. I couldn't make it to OpenStackEast, but I'll be at the midcycle Thurs/Fri.14:04
ildikovsmcginnis: cool! we could almost have a small Cinder gathering there it seems :)14:05
*** dims has joined #openstack-cinder14:05
smcginnisGreat! It will be good to see everyone there.14:07
*** GB21 has joined #openstack-cinder14:07
*** Julien-zte has joined #openstack-cinder14:08
openstackgerritMerged openstack/cinder: Imported Translations from Zanata
*** ebalduf has joined #openstack-cinder14:09
openstackgerritMerged openstack/cinder: Add CI wiki page to driver listing
openstackgerritMerged openstack/cinder: Remove the using of dict.get() in assertIsNone
*** mtanino has quit IRC14:09
ildikovjgriffith: are you there Thurs/Fri as well?14:09
*** sdake has joined #openstack-cinder14:10
openstackgerritArnon Yaari proposed openstack/cinder: New cinder driver to support INFINIDAT InfiniBox
*** savihou has quit IRC14:15
jgriffithildikov: no.. unfortunately I have to head out on Thursday14:17
ildikovjgriffith: ok, never mind, I was just curious14:20
*** jungleboyj has joined #openstack-cinder14:21
*** lpetrut has joined #openstack-cinder14:24
*** Asaithambi has quit IRC14:25
*** ametts has joined #openstack-cinder14:26
*** pcaruana has quit IRC14:30
*** Asaithambi has joined #openstack-cinder14:31
*** jdurgin1 has joined #openstack-cinder14:39
*** Julien-zte has quit IRC14:40
*** JoseMello has quit IRC14:40
*** krtaylor has quit IRC14:42
*** ducttape_ has quit IRC14:44
*** baumann has joined #openstack-cinder14:45
*** charlesr has joined #openstack-cinder14:47
*** sdake has quit IRC14:48
openstackgerritMerged openstack/os-brick: Add connector for vmdk volumes
*** jbernard1 has joined #openstack-cinder14:52
*** xinli has joined #openstack-cinder14:54
*** ociuhandu has joined #openstack-cinder14:55
*** Julien-zte has joined #openstack-cinder14:58
*** akerr is now known as akerr_away15:00
*** akerr_away is now known as akerr15:00
*** lpetrut has quit IRC15:06
*** rcernin has quit IRC15:06
*** mvk has quit IRC15:10
*** Yogi1 has joined #openstack-cinder15:10
*** leeantho has joined #openstack-cinder15:12
*** karthik__ has joined #openstack-cinder15:12
*** Asaithambi has quit IRC15:19
jgriffithhemna: patrickeast smcginnis can any of you give me some more insight to what we're even trying to solve with the locks in os-brick attach/detach and extend?15:26
jgriffithI'm having difficulty justifying their existence :)15:26
nikeshm  : on this new driver patch, owner of patch is showing "Infinidat Cinder CI", is it OK?15:27
hemnawell, I haven't done a lot of testing w/o them and from what I've seen it's just safer15:28
jgriffithhemna: *safer* for what?15:28
hemnato not remove things we shouldn't be re: iscsi sessions, volume paths, etc.15:28
hemnait shouldn't happen either way15:28
hemnaanother thing is running some of the multipath commands to detect multipath devices is slow15:29
smcginnisI thought we actually ran into an issue and that's how we realized the two services were using different locking dirs when we needed some shared ones.15:29
hemnaand running multiples of that at the same time would be bad15:29
hemnasmcginnis, well if we got rid of the locks we wouldn't need the shared lock setting between cinder and nova15:30
hemnaI rewrote the multipath discovery stuff a release or so ago15:30
hemnaso it could relieve some of the concurrency issues15:30
smcginnishemna: But I thought there was a real legitimate case where we needed those locks.15:30
jgriffithhemna: so I think I can solve the removal thing without locks (I don't really see how they solve that anyway)... and as far as multipath discovery being slow... meh15:30
jgriffithhemna: how slow is "slow"15:31
hemnabut what I saw, when you have 100+ multipath volumes on the same host, running multipath takes sometimes upwards of 5 minutes15:31
sdaguehemna: right, so the issue as we dug through it is that shared locks between cinder/nova require a *ton* of work by deployment tools to make it work, which is probably not realistic15:31
*** raj_singh has joined #openstack-cinder15:31
hemnait gets unusable after a certain number of volumes attached to a host15:31
jgriffithhemna: and locks fix that how?15:31
hemnawell, it prevents you from running multiples of that :)15:31
hemnathe new discover code I put helps prevent brick from even calling multipath though15:32
hemnaso, it could not be an issue now15:32
jgriffithhemna: ok, so it's a problem with efficient use of our API's, not really a locking or consistency problem?15:32
*** charlesr has quit IRC15:32
hemnaIt simply hasn't been tested really15:33
*** diogogmt has joined #openstack-cinder15:33
hemnawhat needs to happen is some rally tests with hundreds of volumes attached and see how it goes15:33
hemnait *should* be ok15:33
smcginnisSo if there's not a real legitimate failure that is protected against by locking, let's take out the locks.15:33
*** Julien-zte has quit IRC15:33
smcginnisIs there ever a case where Cinder is doing something while Nova could come in and do something conflicting and vice versa?15:33
jgriffithsmcginnis: that's what I'm saying :)15:33
jgriffithsmcginnis: sdague hemna and I can look at adding my 'safe-to-remove' call if we want15:34
smcginnisjgriffith: Yeah, you're right. :)15:34
hemnasmcginnis, well at the brick level I don't think it matters as much15:34
hemnanova can try and detach while cinder is attaching and doing some actions15:34
jgriffithhemna: not if they use the api's correctly (ie reserve/unreserve)15:34
jgriffithhemna: that's why those calls exist15:35
smcginnisWould we ever have a case where one thing is logging out a session while another one is trying to log in?15:35
hemnajgriffith, yah that's what I'm thinking as well15:35
hemnasmcginnis, yah that can happen15:35
smcginnisjgriffith: The issue may be backends that don't have separate iSCSI target logins per volume.15:35
hemnaif nova is detaching the last volume from a target portal, while you are attaching a new one15:35
hemnabrick would do a logout15:35
jgriffithsmcginnis: ok, but again I don't know how locks solve any of this?15:36
hemnain that case you wouldn't even start the attachment process until after the removal is done15:36
smcginnisjgriffith: I think just an attach and a detach to a backend that has shared iSCSI logins.15:36
hemnainstead of getting half way through the attach process and then get logged out15:36
hemnathe lock would prevent that15:37
*** kfarr has joined #openstack-cinder15:37
jgriffithsmcginnis: ok, so drivers that do that should have a mechanism to inform a caller that they have that scenario no?15:37
hemnawell it's shared iscsi sessions to the same portal15:37
jgriffithhemna: ok, I get that....15:37
smcginnisOr back to connection tracking and seeing there are other volumes using that session so don't log out of it.15:37
jgriffithsmcginnis: +100015:38
hemnabrick has no mechanism for connection tracking though15:38
smcginnisAnd probably shouldn't...15:38
jgriffithsmcginnis: I'm more than happy to try and propose that part of the changes I have up15:38
hemnayah I don't think it should15:38
*** belmoreira has quit IRC15:38
jgriffithI just didn't think there'd be much hope for something like that going in after the first 8 weeks of the release15:38
jgriffithsmcginnis: in fact, I probably have an abandoned patch somewhere that has a chunk of that done already15:39
*** baumann has quit IRC15:39
smcginnisSo when does Nova handling login/logout and when does Cinder? Seems like it should always be one side.15:39
*** laughterwym has quit IRC15:39
jgriffithsmcginnis: either that or Nova should always "ask" cinder if it's ok IMO15:39
openstackgerritTommyLike proposed openstack/python-cinderclient: fix when restore a backup specified by name
*** baumann has joined #openstack-cinder15:39
jgriffithOk, I gotta run for a bit.  I'm happy to dig into this a bit but I'll probably need some input from hemna and smcginnis to make sure I'm covering the cases you're talking about15:40
*** harlowja_at_home has joined #openstack-cinder15:40
johnthetubaguyI am getting confused about the current state, didn't we say nova doing an attach and a detach at the same time, different volumes, but same attachment, would race without that lock?15:40
hemnajohnthetubaguy, I think it would yes15:41
smcginnisjohnthetubaguy: Yeah15:41
johnthetubaguyagreed, long term, we should just ask cinder15:41
*** haplo37__ has joined #openstack-cinder15:41
hemnayou'd see iscsi sessions vanish15:41
johnthetubaguyhemna: yeah, that makes sense15:41
johnthetubaguyfeels like that the thing we need to keep working in newton15:41
hemnafwiw, FC wouldn't have this problem :P15:41
smcginnisI agree it should always only be nova or only be Cinder. The only problem is if one is logging in while the other is logging out.15:41
sdaguethis is because iscsiadm is storing state locally that gets run over?15:41
dansmithsdague: a large part of iscsiadm's job is state15:42
hemnasdague, it's because when we detach the last iscsi volume from a target portal on a host, we do an iscsi logout15:42
dansmithsdague: but I think it's knowing when we should tell iscsiadm to log out after the last LUN is removed15:42
johnthetubaguydansmith: +115:42
sdaguehemna: ok, so what goes wrong if we don't do the logout?15:42
hemnadansmith, that's what it does today15:42
dansmithsdague: we leak connections15:42
hemnawe'll have sessions stuck around that might cause issues later on15:43
dansmithsdague: i.e. we maintain a session connected to the target15:43
dansmithsomething needs to know that nothing else is coming down the pipe,15:43
smcginnisWould we ever have enough iSCSI sessions that would cause problems really?15:43
dansmitheither because we lock,15:43
dansmithor because we are the only one that could be doing a thing15:44
dansmithsmcginnis: I think it comes down to iscsi requirements for sessions, right?15:44
sdaguedansmith: can we leak more than one connection per host, or is this just that every host will always ahve a connection whether or not we need it?15:44
*** rlrossit_ has quit IRC15:44
dansmithsdague: I think it's that you generally login once, separate from actually connecting to targets, and probably some backends won't let you do things until everyone logs out15:45
hemnaif there aren't any paths left from a target portal, we do iscsi logout and flush the state15:45
smcginnisJust speaking for my companies storage, we could have 1 to maybe 5 tops sessions per storage array.15:45
hemnasdague, for multipath, you'd have multiple iscsi sessions (1 per target portal)15:45
dansmithespecially storage stuff like this should *really* be one of those "only one entity should ever be bring up and down sessions"15:45
smcginnishemna: Not always.15:45
smcginnisdansmith: +115:46
hemnawell, you'd have a session per target_portal.  maybe dell's is different.15:46
smcginnishemna: We'll, multiple initiator ports on the compute host to one on the array, but that's just being pedantic so ignore me. :)15:47
hemnaso anyway, at this point I think it's risky as hell to remove the locks without a lot of testing15:48
smcginnisSo I'm still not clear (need to look through the code), when does Cinder manage the sessions and when does nova?15:48
hemnarally tests with hundreds of volumes15:48
hemnacinder and nova don't manage sessions15:49
sdaguehemna: ok, but the locks aren't going to cooperate between cinder / nova15:49
hemnabrick does only based on what it sees on the host15:49
hemnasdague, not unless the lock dir is the same yah15:49
sdaguehemna: no, not even if that15:49
hemnathe iscsi sessions are managed based upon what brick sees on the host.15:50
hemnawhat's left, etc.15:50
dansmithhemna: in the real world nova and cinder can't share locks15:50
hemna^^ above url15:50
*** pbandark has quit IRC15:50
dansmithfor permissions and selinux reasons15:50
dansmithshared directory aside15:50
smcginnishemna: Well, brick is a library for the two projects. So when does brick get triggered to do it by Cinder and when does it get triggered to do it by nova is my question.15:50
hemnacinder triggers it when it does the copy volume <---> image operations and backup15:50
sdagueso we should just assume as a design point sharing locks between projects using the filesystem is not possible15:50
dansmithsdague: +215:51
sdagueso, given that, what's going to go wrong, and what's the mitigation15:51
dansmithit's also fairly counter to "shared nothing" :P15:51
hemnasdague, sure that's probably a safe bet15:51
johnthetubaguysdague: +1 that15:51
*** Apoorva has joined #openstack-cinder15:51
hemnawell, at most you'd get 1 nova attach/detach going at the same time with 1 cinder attach/detach15:51
sdagueand, as a noob, I think I'm hearing that we could get logged out when trying to do an attach15:52
dansmithhemna: you can't logout/disconnect the session if the actual local block device has been opened, right?15:52
hemnabut I think patrickeast has seen problems with that in his CI runs where we get iscsi session failures because of it.15:52
jgriffithmaybe shared target drivers should use a cinder target as a proxy per volume :)15:52
johnthetubaguyhemna: but when they share a single ISCSI target for two connections its worse right?15:52
sdagueso, could we mitigate that with a retry block around attach?15:52
hemnadansmith, it only calls logout if there are no more /dev/disk/by-path/* entries on the same target portal15:52
dansmithright, so there's no data loss problem ehre,15:53
smcginnissdague: Hmm, maybe15:53
hemnabut there is a race between looking for the last device and doing logout15:53
dansmithit's merely a failed boot that thinks it has attached something and then suddenly it's gone, right?15:53
dansmithor a leaked connecton15:53
openstackgerritJon Bernard proposed openstack/cinder: RBD: postpone snapshot removal until final volume deleted
smcginnissdague: It's a fresh attach, so it's not like we'd be yanking the device away while IO is going.15:53
dansmithsmcginnis: right, that's my point15:53
smcginnisdansmith: +115:54
sdaguedansmith: so, what you are saying is that we can't actually corrupt anything, we'll just have the possibility for more failed attaches?15:55
dansmithsdague: right15:55
dansmithsdague: or some potentially confused state15:55
smcginnispatrickeast: ping15:55
dansmithbut nothing disastrous, AFAICT15:55
dansmithwhich is why I think we should document and fix properly15:56
hemnapatrickeast, ping15:56
smcginnishemna: ;)15:56
hemnapatrickeast, he saw some failed CI runs due to the non shared lock15:56
dansmithI'm sure15:56
sdaguedansmith: in the real world, yeh. It might take our gate runs a bit too unreliable with the fast attach/detach at times15:56
smcginnisBut if we add retries around that attach...15:56
* patrickeast yawns15:56
hemnasmcginnis, I'm not sure retries would help here15:56
dansmithsdague: yeah, we could probably retry in the test though right?15:57
patrickeastWhat's up?15:57
hemnabecause the race is in between testing for the last volume and logout15:57
dansmithunless we wedge state15:57
sdaguedansmith: in all volumes tests?15:57
hemnawe could get a volume showing up and passed on to nova between there15:57
*** Apoorva has quit IRC15:57
dansmithsdague: well, write a wrapper, but ...15:57
*** mtanino has joined #openstack-cinder15:57
smcginnispatrickeast: Did you have issues in your CI with attach and detach collisions?15:57
dansmithsdague: either way, some gate-specific fix instead of something that impacts the real world would be more ideal I think15:57
hemnapatrickeast, do you have an examples of the failed CI runs when nova and cinder brick calls collided ?15:58
patrickeastsmcginnis: hemna yea I ran into problems when cranking up tempest concurrency15:58
sdaguedansmith: maybe, we're definitely making volume attach less reliable here, if there was a safe way to reacquire the session if we lost it in the middle of attach that would be nice15:58
hemnawhich is what I'd expect15:58
patrickeastI can see if the links are still alive, they are only archived for like 30 days or something15:59
dansmithsdague: the whole point of this is that there isn't without a lock or a single thread15:59
dansmithsdague: what is the thing that is causing this in the gate that isn't there now? maybe I'm missing that15:59
* smcginnis jumps on a call and will be half paying attention15:59
patrickeastBut yea, reading scrollback, it's just failed attachments15:59
dansmithsdague: sounds like you mean something new is being added15:59
sdaguedansmith: so if you run iscsiadm after the session logout happened, will we get a distinctive error16:00
hemnamaybe if we added a retry into the brick connect_volume call16:00
dansmithsdague: I dunno, maybe.. at least on stderr16:00
sdagueI wonder if we could treat this like keystone expired tokens16:01
sdagueif you run a command, get a "no session" error, just force creating a new session and try that command again16:01
dansmiththere's a bunch of detailed error codes for iscsiadm, so probably16:01
sdaguebut, I'm a noob in this space, so I don't know if that would work or be a thing16:01
dansmithsdague: the question is whether we have the login information at the point at which we need to re-login to a thing we're just connecting16:01
dansmithprobably this for a situation where we've already logged out: ISCSI_ERR_NO_OBJS_FOUND16:02
patrickeastYea something like that seems like it would probably be ok16:02
* patrickeast wanders off to start getting ready for work16:03
*** GB21 has quit IRC16:05
openstackgerritMerged openstack/cinder: Remove unused context parameter
openstackgerritJustin A Wilson proposed openstack/cinder: Added config option to enable SSL
sdagueok... so where do we stand? do we think making attach more robust to the logging out behind the scenes is the right path? And if so, who's got the expertise and time to do that thing before freeze?16:09
*** dims has quit IRC16:10
dansmithsmcginnis: hemna: looking at I don't see why you couldn't make that code retry and relogin if we failed to attach because something else logged us out16:11
smcginnisSorry, too distracted, but I think you're probably right dansmith.16:12
smcginnisWe can look at doing that.16:12
hemnawell the race could happen if a connect_volume call is returning and we are also doing an iscsi logout16:14
hemnait's pretty much too late then w/o retrying connect_volume again16:14
dansmithhemna: you said that the logout code wouldn't run if /dev/disk/by-id/$foo exists, right?16:14
dansmithhemna: so by the time connect_volume returns, you know that it's locked in place, yes?16:14
hemnayah, but we could test for the paths and get none, and then have a login happen before logout gets called16:15
hemnacreating the race16:15
hemnanova would get a volume path that's broken at that point16:15
sdaguehemna: how long is that window?16:15
hemnait's short16:15
hemnavery short16:15
hemnabut possible on a busy host16:15
dansmithshorter than the already short one16:16
openstackgerritMerged openstack/cinder: [doc] Fix a non-alphabetical order bug
openstackgerritMerged openstack/cinder: Remove deprecaterd mapping in
sdagueI think we just document that as a known issue16:16
hemnaI think this one would just be a retry16:16
hemnaand could even be user driven16:16
dansmithsdague: it's really the same as today I think16:17
*** e0ne has quit IRC16:17
dansmithsdague: a retry closes the window a little, but it's still the same user experience, AFAIK16:17
hemnaw/o the shared lock dir, today the window is between 2 concurrent attach/detach calls (1 for nova and 1 for cinder)16:17
hemnawithout any locks, that window is N16:17
*** ducttape_ has joined #openstack-cinder16:18
hemnaI could put up a patch removing the locks and we could see how it goes in CI :P16:18
hemnacrank up the concurrency in the 3prd party CI runs16:18
sdaguedansmith: well, any mitigation would be good16:18
hemnaalso sounds like we could use a rally test for brick too16:19
sdagueand maybe there would be a way to reorder the code to get past that some other way16:19
dansmithsdague: I don't see any ordering fix.. I was thinking about that.. I think because of the way iscsiadm is16:19
*** karthik__ has quit IRC16:19
*** Suyi has joined #openstack-cinder16:22
sdaguedansmith: ok, so even if in here you did another device check before the logout, and one after (and if the device check after failed, you recreate the session?) -
sdagueanyway, any ways we can narrow this window would be good, and if testing becomes too unreliable, we look into other mitigations16:25
dansmithsdague: the problem is you still can have some process somewhere blocked on running iscsiadm --logout, and if it runs after you, you're toast16:26
sdaguedansmith: sure16:26
dansmithso my point earlier was narrow the window, but can't close it without more work16:26
hemnaI'll put up a WIP that removes the locks and we can see how the runs go16:27
hemnathe only problem is that the traffic is really light in terms of the number of volumes at attach/detach work16:27
hemnaa rally test would be more suited for it16:28
*** harlowja_at_home has quit IRC16:28
*** ametts_ has joined #openstack-cinder16:28
sdaguehemna: ok, though, that is going to potentially allow n-cpu to race with itself16:30
*** ametts has quit IRC16:30
hemnawell n-cpu with c-vol16:30
*** yangyapeng has quit IRC16:31
*** dkehn_ has quit IRC16:31
sdagueno, n-cpu with itself right, because we're in eventlet, and if we switch up greenthreads here by doing io we could be processing an attach and detach at the same time, right?16:31
hemnathat's one case yah16:31
*** Apoorva has joined #openstack-cinder16:31
*** xinli has quit IRC16:31
sdaguehemna: so I don't think that's a useful exercise16:32
hemnathe real issue though is the iscsi sessions16:32
sdagueI think narrowing the race window as much as possible for newton is the thing to do16:32
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: Mask out passwords when tracing
*** bwallis has quit IRC16:36
*** cknight has joined #openstack-cinder16:36
*** daneyon_ has joined #openstack-cinder16:38
*** baumann_ has joined #openstack-cinder16:38
*** daneyon has quit IRC16:40
*** diablo_rojo has quit IRC16:41
*** baumann has quit IRC16:41
*** dims has joined #openstack-cinder16:42
_alastor_If anyone has a moment, I'd appreciate some eyes on an update for my driver:
*** dims has quit IRC16:47
*** dkehn has joined #openstack-cinder16:48
*** alonma has quit IRC16:51
*** alonma has joined #openstack-cinder16:51
*** dims has joined #openstack-cinder16:53
*** GB21 has joined #openstack-cinder16:55
*** alonma has quit IRC16:56
openstackgerritMerged openstack/cinder: Replication, hard-coding and issues in K2
openstackgerritOpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements
*** alonma has joined #openstack-cinder16:59
*** sdake has joined #openstack-cinder17:00
*** jungleboyj has quit IRC17:00
*** bwallis has joined #openstack-cinder17:02
*** daneyon has joined #openstack-cinder17:03
*** alonma has quit IRC17:03
*** alonma has joined #openstack-cinder17:06
*** alonma has quit IRC17:06
*** daneyon_ has quit IRC17:06
*** alonma has joined #openstack-cinder17:06
*** daneyon_ has joined #openstack-cinder17:07
*** harlowja has joined #openstack-cinder17:08
*** daneyon has quit IRC17:09
*** rajinir has joined #openstack-cinder17:12
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: WIP remove iscsi and fibre channel locks
*** jungleboyj has joined #openstack-cinder17:13
openstackgerritArnon Yaari proposed openstack/cinder: New cinder driver to support INFINIDAT InfiniBox
*** yuelongguang has quit IRC17:21
*** krtaylor has joined #openstack-cinder17:22
*** daneyon has joined #openstack-cinder17:22
*** Lee1092 has quit IRC17:23
*** yuelongguang has joined #openstack-cinder17:23
*** johnplanow has joined #openstack-cinder17:24
*** daneyon_ has quit IRC17:26
*** bardia has joined #openstack-cinder17:29
*** bardia has quit IRC17:32
*** bardia has joined #openstack-cinder17:32
*** rcernin has joined #openstack-cinder17:37
*** kfarr has quit IRC17:37
*** e0ne has joined #openstack-cinder17:38
*** diablo_rojo has joined #openstack-cinder17:42
*** JoseMello has joined #openstack-cinder17:47
openstackgerritAlon Marx proposed openstack/cinder: Fix invalid services caused by enabled_backends
*** raunak has joined #openstack-cinder17:48
*** xinli has joined #openstack-cinder17:49
*** raj_singh has left #openstack-cinder17:50
openstackgerritMatt Smith (_alastor_) proposed openstack/cinder: Datera driver 2.2 update
openstackgerritxing-yang proposed openstack/cinder: WIP: Migrate consistency groups to groups
*** GB21 has quit IRC17:56
*** jsheeren has joined #openstack-cinder18:02
*** ankur-gupta-f has joined #openstack-cinder18:06
ankur-gupta-fthrawn01: fyi18:07
*** e0ne has quit IRC18:15
*** e0ne has joined #openstack-cinder18:17
*** Apoorva_ has joined #openstack-cinder18:20
*** xyang1 has quit IRC18:22
*** Apoorva has quit IRC18:24
*** xyang1 has joined #openstack-cinder18:26
*** raj_singh has joined #openstack-cinder18:29
*** neophy has joined #openstack-cinder18:33
*** rcernin has quit IRC18:42
*** Apoorva_ has quit IRC18:46
*** karthik__ has joined #openstack-cinder18:46
*** Apoorva has joined #openstack-cinder18:47
*** jsheeren has quit IRC18:48
patrickeastscottda: ping18:49
*** haplo37__ has quit IRC18:50
*** e0ne has quit IRC18:59
scottdapatrickeast: Hi19:00
hemnaso much pain19:01
hemnadrivers that don't extend the BaseVD or VolumeDriver19:01
patrickeastscottda: hey, so i saw merge, was that the last blocker before we can push on the tempest tests for retype/migration?19:01
hemnaP A I N19:01
hemnashoot me19:01
patrickeasthemna: lol19:01
patrickeastscottda: or i guess first?19:01
* patrickeast doesn't know what order things need to go in19:01
patrickeastscottda: some context for why i ask, I'm trying to decide if I just wait for things upstream or if I pull the tests and stuff into my internal fork and spend time to try and get it working earlier19:03
scottdapatrickeast: Cool, that's good news...Yeah, that ^^^ needs a rebase and then I'll ping QA to re-apply the +A they already gave...19:03
scottdaWell, I'm rebasing now, so hopefully that ^^ patch get's in pretty quickly...19:04
scottdaalready reviewed and approved and all..19:04
*** haplo37__ has joined #openstack-cinder19:04
patrickeastthen and  should be good to go, right?19:04
* hemna wonders how many drivers he's going to break today........19:05
patrickeasthemna: are you changing base classes for drivers?19:05
scottdapatrickeast: Yes, although I'm pushing first on (retype with volume attached)..19:05
* hemna hides19:06
patrickeastscottda: sounds good19:06
nikeshmhi what is the command to check python3 compatabilty for our cinder driver files19:06
hemnarm -rf19:06
scottdapatrickeast: Did gate-grenade-dsvm-cinder-multinode-nv get merged in? That'd be cool...19:06
patrickeastscottda: uhh i think so19:07
nikeshmalso this cinder driver patch owner name is CI name --- -- is it OK?19:07
*** fifieldt has quit IRC19:07
patrickeastscottda: although i dunno where the link is to check, lol19:07
scottdapatrickeast: Yeah, sorry, it looks like it did from the merge conflict. That was less a question and more of a way-of-speaking thing...19:07
nikeshmsome command was :  tox -...............19:08
*** asselin_ has joined #openstack-cinder19:08
patrickeastnikeshm: just run tox -epy3419:08
nikeshmthanks.....this cinder driver patch owner name is CI name --- -- is it OK?19:09
scottdapatrickeast: OK, rebase of 330678 is up. When it passes jenkins, I'll see if I can get it re-reviewed...19:10
*** haplo37__ has quit IRC19:10
patrickeastscottda: sweet19:10
*** asselin has quit IRC19:10
*** kfarr has joined #openstack-cinder19:10
smcginnisnikeshm: Thanks for pointing that out.19:10
patrickeastnikeshm: i don't think it really matters who the patch owner is19:10
scottdapatrickeast: BTW, have you had a chance to try Gorka's manual tests for HA? They worked for me, using cleanup and job distribution.19:11
patrickeastscottda: not yet :( one of these days I'll get around to it19:11
*** ametts_ has quit IRC19:13
nikeshmpatrickeast: had concern from this
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP Add new supported driver checks
patrickeasthemna: ouch19:15
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP Add new supported driver checks
hemnapatrickeast, yup19:16
* hemna runs!19:16
*** ametts_ has joined #openstack-cinder19:18
smcginnishemna: I'd rather just set the default that everything is supported. Then if we have a non-compliant driver we can just put up a patch to have a one line change in that driver saying supported=False.19:18
*** ametts_ has quit IRC19:18
patrickeastsmcginnis: +119:19
*** fifieldt has joined #openstack-cinder19:19
hemnaheh ok19:19
*** ametts has joined #openstack-cinder19:19
smcginnishemna: Now that you've done all that typing. :P19:19
hemnamy hands hurt too much to update it right now....I need fewd.19:19
jungleboyjMmmm, fewd.19:20
smcginnisfewd es gud19:20
jungleboyjI luv fewd.19:21
scottdaDid somebody need me? Oh, just my automatic Food ping....19:21
*** haplo37__ has joined #openstack-cinder19:22
*** asselin__ has joined #openstack-cinder19:25
*** asselin_ has quit IRC19:28
*** baumann_ has quit IRC19:41
*** baumann has joined #openstack-cinder19:42
*** daneyon_ has joined #openstack-cinder19:55
*** daneyon has quit IRC19:58
*** alonma has quit IRC20:00
*** Apoorva has quit IRC20:06
*** liverpooler has quit IRC20:07
*** alonma has joined #openstack-cinder20:08
*** ankur-gupta-f has left #openstack-cinder20:12
*** enriquetaso has joined #openstack-cinder20:12
*** alonma has quit IRC20:13
*** neophy has quit IRC20:14
hemnascottda, damn I don't have an automatic feeder.....20:15
*** alonma has joined #openstack-cinder20:16
*** porrua has quit IRC20:16
*** Apoorva has joined #openstack-cinder20:17
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP Add new supported driver checks
*** Yogi1 has quit IRC20:19
*** enriquetaso has quit IRC20:19
hemnaoooh....can I fill that with burritos?20:20
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP Add new supported driver checks
*** alonma has quit IRC20:20
jungleboyjsmcginnis: Nice, I could have used one of those for our cat.20:20
*** karthik__ has quit IRC20:21
SwansonThe comments "can I fill that with burritos?" and "I could have used one of those for our cat" do not appear to refer to a common starting point.20:21
hemnaSwanson, I'm always trying to stuff a square peg into a round hole20:22
jungleboyjSwanson: Yeah, I don't think hemna has/had a cat.20:22
hemnatraumatized as a child20:22
* jungleboyj pictures a feeder shooting burritos as hemna 20:22
*** akerr has quit IRC20:23
*** cargonza has joined #openstack-cinder20:23
patrickeastwait, you guys saying I can't feed a cat burritos?20:23
*** baumann_ has joined #openstack-cinder20:23
jungleboyjpatrickeast: Only if you want to clean up the resulting mess.20:23
SwansonMy cats are 15 yo. They shoot burritos no matter what you feed them.20:24
jungleboyjOy vey20:24
SwansonThe things they did to their litter box when I went on vacation was diabolical.20:26
*** baumann has quit IRC20:26
openstackgerritWalter A. Boring IV (hemna) proposed openstack/os-brick: WIP remove iscsi and fibre channel locks
SwansonEasier pets to deal with --
*** asselin__ has quit IRC20:28
smcginnisw. t. f!20:29
*** asselin has joined #openstack-cinder20:29
*** xyang1 has quit IRC20:29
jungleboyjI don't need to buy those.  They are all over my house .... hmmm, can I just make a living selling those?20:29
*** xyang1 has joined #openstack-cinder20:29
SwansonThey're the perfect gift.20:29
smcginnisAt $37.95 for 12, you might have a good side business there. ;)20:29
smcginnis"Not sexed" Imagine having the job of checking and sorting the spiders.20:30
hemnaprice is a little high20:30
hemnamight be able to find them cheaper on eBay from China20:30
hemnafree shipping too.20:30
*** ociuhandu has quit IRC20:31
smcginnisThis site is crazy! I should have known there was somewhere inline to buy a batch of pill bugs, but man!20:31
*** sdake has quit IRC20:32
*** asselin_ has joined #openstack-cinder20:32
smcginnisOoh, virus kits.20:32
*** xyang has joined #openstack-cinder20:32
*** yangyapeng has joined #openstack-cinder20:33
hemnaooh, and I'm fresh out of e-coli20:34
*** asselin has quit IRC20:35
*** alonma has joined #openstack-cinder20:39
* jungleboyj wonders why Swanson knew about that site.20:41
smcginnisjungleboyj: I've found it's better not to ask.20:42
*** alonma has quit IRC20:44
* jungleboyj looks the other direction.20:44
SwansonIt was in my twitter feed next to this ...
*** alonma has joined #openstack-cinder20:47
jungleboyjI almost lost my lunch there.  Thanks Swanson.20:48
SwansonOh, don't scroll down.20:48
jungleboyjOh jeez.20:50
*** gouthamr has quit IRC20:51
*** alonma has quit IRC20:52
Swansonjungleboyj, was that the lunch?20:54
*** alonma has joined #openstack-cinder20:54
jungleboyjSwanson: Thankfully I am good at controlling my stomach.20:54
johnplanowThe time I've spent in openstack-cinder channel recently? It just paid off. ;)20:55
hemnajohnplanow, it's all about the diversity.20:55
jungleboyjjohnplanow: Always some ROI here.20:55
johnplanowI've learned more in the past 30 mins..20:56
*** sdake has joined #openstack-cinder20:57
*** alonma has quit IRC20:59
SwansonI try to give back.20:59
xyangpatrickeast: ping21:00
patrickeastxyang: yo21:00
*** JoseMello has quit IRC21:00
xyangpatrickeast: thanks for submitting the CG tempest tests21:00
xyangpatrickeast: they were not run by your CI, is that dependent on the devstack patch21:00
patrickeastxyang: yea, there isn't a good way to update policy.json right now21:01
xyangpatrickeast: ok, have you tried it on your setup manually21:01
patrickeastxyang: yep, works fine as long as the cg api's are available21:02
xyangpatrickeast: so we need to wait for the devstack patch to get in first?  that will take a while, right21:02
patrickeastxyang: why?21:02
patrickeastxyang: need that in for CI systems to use it... but anyone can run tempest with the cinder plugin before then21:03
*** lpetrut has joined #openstack-cinder21:03
*** haplo37__ has quit IRC21:03
*** alonma has joined #openstack-cinder21:03
*** julim has quit IRC21:03
xyangpatrickeast: ok, I don't have a problem to get it in first21:03
patrickeastxyang: yay :D21:04
xyangpatrickeast: it's just that I can only do a +1 now since my name is also on it:)21:06
*** alonma has quit IRC21:08
*** sdake has quit IRC21:08
*** alonma has joined #openstack-cinder21:10
*** lpetrut has quit IRC21:12
*** alonma has quit IRC21:15
*** dustins has quit IRC21:15
*** sdake has joined #openstack-cinder21:17
*** pauloewerton has quit IRC21:18
*** alonma has joined #openstack-cinder21:18
*** haplo37__ has joined #openstack-cinder21:18
patrickeastugh release notes21:19
*** mvk has joined #openstack-cinder21:19
*** alonma has quit IRC21:22
*** xinli has quit IRC21:22
*** lpetrut has joined #openstack-cinder21:23
*** alonma has joined #openstack-cinder21:25
openstackgerritPatrick East proposed openstack/cinder: Add get_manageable_* methods to Pure drivers
openstackgerritTom Barron proposed openstack/cinder: Use constraints for all tox environments
*** ametts has quit IRC21:28
*** alonma has quit IRC21:30
*** baumann_ has quit IRC21:30
*** alonma has joined #openstack-cinder21:32
*** diablo_rojo has quit IRC21:34
*** jungleboyj has quit IRC21:34
*** alonma has quit IRC21:37
*** xinli has joined #openstack-cinder21:38
*** sdake has quit IRC21:39
*** alonma has joined #openstack-cinder21:40
openstackgerritWalter A. Boring IV (hemna) proposed openstack/cinder: WIP Add new supported driver checks
*** alonma has quit IRC21:44
*** alonma has joined #openstack-cinder21:47
*** raj_singh has left #openstack-cinder21:47
openstackgerritxing-yang proposed openstack/cinder: Add generic volume groups
*** alonma has quit IRC21:52
*** sdake has joined #openstack-cinder21:54
*** alonma has joined #openstack-cinder21:54
openstackgerritxing-yang proposed openstack/cinder: Add group snapshots
*** harlowja has quit IRC21:55
*** alonma has quit IRC21:59
*** johnplanow has quit IRC22:00
*** alonma has joined #openstack-cinder22:02
*** mtanino has quit IRC22:03
*** xyang has quit IRC22:06
*** xyang has joined #openstack-cinder22:07
*** alonma has quit IRC22:07
*** mriedem has quit IRC22:09
*** alonma has joined #openstack-cinder22:10
*** kfarr has quit IRC22:10
*** lpetrut has quit IRC22:11
jamielennoxhey cinder, i've had up for a while reworking how you load context objects. can i get some eyes please?22:13
*** alonma has quit IRC22:14
*** cknight has quit IRC22:14
*** xinli has quit IRC22:14
*** haplo37__ has quit IRC22:15
*** bswartz has quit IRC22:15
jgriffithjamielennox: love to help, but man that patch is kinda huge22:17
*** jistr has quit IRC22:18
*** xyang has quit IRC22:18
jamielennoxi had to double check there i had the right patch linked before realizing the sarcasm22:19
jgriffithjamielennox: :)22:19
jgriffithjamielennox: happy Monday!22:19
*** jistr has joined #openstack-cinder22:19
smcginnisjamielennox: I like that - +23, -4122:20
*** sdake has quit IRC22:20
jgriffithjamielennox: You may notice that any patch that has an overall reduction in LOC I tend to review more happily22:21
jamielennoxactually looking at it that patch was a while ago and i should probably be able to get another -5 out with recent release, but we can start with that22:21
jamielennoxjgriffith: yep, also hopefully anything that removes oslo boilerplate code22:21
jgriffithjamielennox: I should make that a filter... some just look for the opportunity to add the final +2/A... me, I'll just look for code reduction22:21
*** xyang has joined #openstack-cinder22:21
smcginnisNot a bad plan.22:22
*** xyang has quit IRC22:22
*** xyang has joined #openstack-cinder22:22
*** jamespd has joined #openstack-cinder22:25
jamespdIf I am seeing "Deadlock detected when running 'quota_reserve'" in nova-volume/8.0.0, should I be submitting that as a bug?22:26
jamespdwhen users delete a bunch of volumes at once, cinder-volume hits the deadlocks and everything is sad.22:27
* jamespd would love some advice.22:27
jgriffithjamespd: certainly sounds like a bug to me.  Would be great if you posted it with some log files if you could and detail on how to repro22:28
*** xyang has quit IRC22:28
*** sdake has joined #openstack-cinder22:35
openstackgerritAbhilash Divakaran proposed openstack/cinder: Add CI_WIKI_NAME to Tegile volume driver
jamespdjgriffith: thanks!22:36
jamespds/nova-volume/cinder-volume/, obviously.22:36
openstackgerritAbhilash Divakaran proposed openstack/cinder: Add CI_WIKI_NAME to Tegile volume driver
openstackgerritAbhilash Divakaran proposed openstack/cinder: Add CI_WIKI_NAME to Tegile volume driver
*** krtaylor has quit IRC22:38
*** ducttape_ has quit IRC22:40
openstackgerritMatt Smith (_alastor_) proposed openstack/cinder: Datera driver 2.2 update
scottdasmcginnis: IIRC, we're not ready to deprecate anything in the cinderclient (in favor of openstack client), so we don't want this, right?  Wanted to check before -1'ing...or -2'ing..22:45
*** edmondsw has quit IRC22:51
*** bardia has quit IRC22:53
*** bardia has joined #openstack-cinder22:53
*** sdake has quit IRC22:57
*** mriedem has joined #openstack-cinder22:57
*** markvoelker has quit IRC23:06
*** xyang1 has quit IRC23:08
*** sticker has joined #openstack-cinder23:08
*** asselin has joined #openstack-cinder23:14
*** asselin_ has quit IRC23:15
*** asselin_ has joined #openstack-cinder23:15
*** asselin has quit IRC23:19
*** alonma has joined #openstack-cinder23:31
*** diogogmt has quit IRC23:31
*** bwallis has quit IRC23:33
*** alonma has quit IRC23:35
*** harlowja has joined #openstack-cinder23:36
*** alonma has joined #openstack-cinder23:41
*** sdague has quit IRC23:44
*** alonma has quit IRC23:45
*** alonma has joined #openstack-cinder23:48
*** leeantho has quit IRC23:52
*** alonma has quit IRC23:53
*** bwallis has joined #openstack-cinder23:55
*** asselin_ has quit IRC23:56

Generated by 2.14.0 by Marius Gedminas - find it at!