Friday, 2017-08-11

*** thingee_ has quit IRC00:00
*** baojg has quit IRC00:02
*** itlinux has quit IRC00:06
mriedemthere are some volume group tests that have been failing randomly since they were added a day or so ago
openstackLaunchpad bug 1709739 in Cinder ""Invalid volume: Volume is not local to this node" in neutron multinode job (pike)" [Undecided,Confirmed]00:12
mriedem68 hits in 2 days00:12
*** baojg has joined #openstack-cinder00:14
*** baojg has quit IRC00:14
*** baojg has joined #openstack-cinder00:15
*** baojg has quit IRC00:15
*** k2so is now known as sdake00:16
*** edmondsw has joined #openstack-cinder00:17
*** yangyapeng has quit IRC00:18
*** sbezverk has quit IRC00:21
*** edmondsw has quit IRC00:22
openstackgerritMerged openstack/cinder master: Revert "Mark ETERNUS drivers as unsupported"
*** zhurong has joined #openstack-cinder00:51
*** markvoelker has joined #openstack-cinder01:03
*** itlinux has joined #openstack-cinder01:04
*** Shunli has joined #openstack-cinder01:04
*** Apoorva_ has quit IRC01:09
*** baojg has joined #openstack-cinder01:16
*** yangyapeng has joined #openstack-cinder01:21
*** baojg has quit IRC01:24
*** PsionTheory has quit IRC01:35
*** edmondsw has joined #openstack-cinder02:06
*** edmondsw has quit IRC02:10
*** hemna__ has joined #openstack-cinder02:11
*** zsli_ has joined #openstack-cinder02:11
*** yangyape_ has joined #openstack-cinder02:11
*** hemna_ has quit IRC02:11
*** diablo_rojo has quit IRC02:17
*** yangyapeng has quit IRC02:18
*** Shunli has quit IRC02:18
*** tomh has quit IRC02:18
*** tomh has joined #openstack-cinder02:18
*** jistr has quit IRC02:18
*** geguileo has quit IRC02:18
*** openstackgerrit has quit IRC02:21
*** armax has quit IRC02:21
*** lifeless_ has joined #openstack-cinder02:21
*** baojg has joined #openstack-cinder02:21
*** armax has joined #openstack-cinder02:21
*** jistr has joined #openstack-cinder02:23
*** tomh_ has joined #openstack-cinder02:23
*** baojg has quit IRC02:26
*** geguileo has joined #openstack-cinder02:27
*** wanghao has quit IRC02:27
*** wanghao has joined #openstack-cinder02:28
*** lifeless has quit IRC02:28
*** tomh has quit IRC02:28
*** openstackgerrit has joined #openstack-cinder02:36
openstackgerritYu Zhang proposed openstack/cinder master: Storwize:Update document for Pike release
openstackgerritOpenStack Release Bot proposed openstack/cinder master: Update reno for stable/pike
openstackgerritYu Zhang proposed openstack/cinder master: Storwize:Update document for Pike release
*** dave-mccowan has quit IRC03:06
*** nicolasbock has joined #openstack-cinder03:21
*** mriedem has quit IRC03:22
*** sbezverk has joined #openstack-cinder03:24
*** mriedem has joined #openstack-cinder03:27
*** psachin has joined #openstack-cinder03:41
*** psachin has quit IRC03:43
*** itlinux_ has joined #openstack-cinder03:43
openstackgerrityfzhao proposed openstack/cinder master: Fix wrong links
*** psachin has joined #openstack-cinder03:51
*** yangyape_ has quit IRC03:51
*** edmondsw has joined #openstack-cinder03:54
*** fredli__ has joined #openstack-cinder03:56
*** mvk has joined #openstack-cinder03:57
*** edmondsw has quit IRC03:58
*** chhavi has joined #openstack-cinder04:00
*** itlinux_ has quit IRC04:03
*** sbezverk has quit IRC04:07
*** chaynikasaikia has joined #openstack-cinder04:09
openstackgerrityfzhao proposed openstack/cinder master: Fix wrong links
*** itlinux_ has joined #openstack-cinder04:16
*** itlinux_ has quit IRC04:19
*** chaynikasaikia has quit IRC04:19
*** itlinux_ has joined #openstack-cinder04:19
*** itlinux_ has quit IRC04:23
*** baojg has joined #openstack-cinder04:23
*** baojg has quit IRC04:30
*** adisky__ has joined #openstack-cinder04:31
*** markstur has quit IRC04:32
*** chaynikasaikia has joined #openstack-cinder04:43
*** chaynikasaikia has left #openstack-cinder04:43
*** abhitechie has joined #openstack-cinder04:44
*** abhitechie has quit IRC04:45
*** abhitechie has joined #openstack-cinder04:45
*** savvas has joined #openstack-cinder04:50
*** savvas has joined #openstack-cinder04:51
*** Poornima has joined #openstack-cinder04:51
*** Poornima_K has joined #openstack-cinder05:17
*** Poornima has quit IRC05:18
*** savvas has quit IRC05:19
*** markstur has joined #openstack-cinder05:35
*** markstur has quit IRC05:40
*** Poornima_K has quit IRC05:40
*** Poornima has joined #openstack-cinder05:40
openstackgerritMerged openstack/cinder master: Update reno for stable/pike
*** edmondsw has joined #openstack-cinder05:42
*** edmondsw has quit IRC05:46
*** armax has quit IRC05:51
*** lpetrut has joined #openstack-cinder05:52
*** armax has joined #openstack-cinder05:52
*** armax has quit IRC05:52
*** armax has joined #openstack-cinder05:52
*** armax has quit IRC05:53
*** armax has joined #openstack-cinder05:53
*** armax has quit IRC05:54
*** armax has joined #openstack-cinder05:54
*** armax has quit IRC05:54
*** itlinux has quit IRC05:55
*** armax has joined #openstack-cinder05:55
*** armax has joined #openstack-cinder05:56
*** armax has quit IRC05:56
*** armax has joined #openstack-cinder05:57
*** itlinux has joined #openstack-cinder05:57
*** armax has joined #openstack-cinder05:57
*** armax has quit IRC05:58
*** fredli__ has quit IRC06:05
*** _pewp_ has quit IRC06:13
*** _pewp_ has joined #openstack-cinder06:14
*** tesseract has joined #openstack-cinder06:16
*** rcernin has joined #openstack-cinder06:22
*** ChubYann has quit IRC06:26
*** baojg has joined #openstack-cinder06:27
*** itlinux has quit IRC06:28
*** savvas has joined #openstack-cinder06:31
*** rtjure has joined #openstack-cinder06:32
*** baojg has quit IRC06:34
*** savvas has quit IRC06:35
*** Poornima has quit IRC06:37
*** libin_ has joined #openstack-cinder06:43
*** Poornima has joined #openstack-cinder06:52
*** lpetrut has quit IRC06:52
*** pgadiya has joined #openstack-cinder07:03
*** aarefiev_afk is now known as aarefiev07:08
*** Poornima has quit IRC07:10
*** saphi has joined #openstack-cinder07:13
*** baojg has joined #openstack-cinder07:30
*** baojg has quit IRC07:37
*** e0ne has joined #openstack-cinder07:42
*** Poornima has joined #openstack-cinder07:52
*** openstackgerrit has quit IRC08:02
*** mvk has quit IRC08:14
*** openstackgerrit has joined #openstack-cinder08:20
openstackgerritYingzhe Zeng proposed openstack/cinder master: Fix to get correct hypermetro remote lun ID
*** mvk has joined #openstack-cinder08:21
*** yangyapeng has joined #openstack-cinder08:30
*** Lee1092 has joined #openstack-cinder08:32
openstackgerrityixuan zhang proposed openstack/cinder master: Storwize: add hyperswap volume support
*** gfidente has joined #openstack-cinder08:40
*** gfidente has left #openstack-cinder08:40
*** wanghao has quit IRC08:40
*** wanghao has joined #openstack-cinder08:41
*** wanghao_ has joined #openstack-cinder08:44
*** wanghao has quit IRC08:48
openstackgerrityfzhao proposed openstack/python-cinderclient master: Fix wrong links
*** yangyapeng has quit IRC09:03
*** wanghao has joined #openstack-cinder09:09
*** chhavi has quit IRC09:10
*** wanghao_ has quit IRC09:12
*** markvoelker has quit IRC09:17
*** wanghao has quit IRC09:24
*** wanghao has joined #openstack-cinder09:26
*** mriedem has quit IRC09:30
*** baojg has joined #openstack-cinder09:34
*** zsli_ has quit IRC09:37
*** baojg has quit IRC09:40
*** dalgaaf has quit IRC09:42
*** dalgaaf has joined #openstack-cinder09:44
*** sdague has joined #openstack-cinder10:00
*** chhavi has joined #openstack-cinder10:00
*** pgadiya has quit IRC10:01
*** yangyapeng has joined #openstack-cinder10:04
*** sapd has joined #openstack-cinder10:07
*** ociuhandu has quit IRC10:08
*** yangyapeng has quit IRC10:09
*** pgadiya has joined #openstack-cinder10:18
*** markvoelker has joined #openstack-cinder10:18
*** alkhodos has quit IRC10:19
*** gkadam has joined #openstack-cinder10:20
*** psachin has quit IRC10:22
*** markvoelker has quit IRC10:23
*** wanghao has quit IRC10:23
*** markstur has joined #openstack-cinder10:31
*** psachin has joined #openstack-cinder10:35
*** markstur has quit IRC10:36
*** markvoelker has joined #openstack-cinder10:39
*** markvoelker_ has joined #openstack-cinder10:44
*** markvoelker has quit IRC10:44
*** markvoelker_ has quit IRC10:44
*** markvoelker has joined #openstack-cinder10:45
*** saphi has quit IRC10:59
*** zhurong has quit IRC11:00
*** yangyapeng has joined #openstack-cinder11:05
*** mriedem has joined #openstack-cinder11:07
*** edmondsw has joined #openstack-cinder11:10
*** yangyapeng has quit IRC11:10
*** sapd has quit IRC11:14
*** dalgaaf has quit IRC11:24
*** e0ne has quit IRC11:24
*** sapd has joined #openstack-cinder11:28
*** dalgaaf has joined #openstack-cinder11:29
*** sbezverk has joined #openstack-cinder11:42
*** liverpooler has joined #openstack-cinder11:45
*** abhitechie has quit IRC11:47
*** sapd has quit IRC11:48
*** dikonoor has joined #openstack-cinder11:49
*** sbezverk has quit IRC11:50
*** e0ne has joined #openstack-cinder11:51
*** yuchenghui has joined #openstack-cinder11:52
*** yuchenghui has quit IRC11:58
*** sapd has joined #openstack-cinder12:01
*** yuchenghui has joined #openstack-cinder12:01
*** _heena_py_ has joined #openstack-cinder12:03
*** yuchenghui has quit IRC12:07
*** markvoelker_ has joined #openstack-cinder12:10
*** markvoelker has quit IRC12:13
*** dikonoor has quit IRC12:15
*** dikonoor has joined #openstack-cinder12:16
*** pgadiya has quit IRC12:16
*** phealy has joined #openstack-cinder12:16
*** zeestrat has quit IRC12:19
*** TheJulia has quit IRC12:19
*** guyr-infinidat has quit IRC12:19
*** TheJulia has joined #openstack-cinder12:19
*** guyr-infinidat has joined #openstack-cinder12:19
*** gouthamr has joined #openstack-cinder12:19
*** zeestrat has joined #openstack-cinder12:20
*** lpetrut has joined #openstack-cinder12:22
*** rtjure has quit IRC12:23
*** dave-mccowan has joined #openstack-cinder12:25
_heena_py_does anyone know how to get volume attachments in python12:25
*** libin_ has quit IRC12:25
*** rtjure has joined #openstack-cinder12:26
*** libin_ has joined #openstack-cinder12:26
*** xyang1 has joined #openstack-cinder12:29
*** catintheroof has joined #openstack-cinder12:30
*** gkadam has quit IRC12:37
*** abishop has joined #openstack-cinder12:39
*** dustins has joined #openstack-cinder12:54
*** savvas has joined #openstack-cinder13:01
*** baojg has joined #openstack-cinder13:03
*** yangyapeng has joined #openstack-cinder13:07
*** Poornima has quit IRC13:07
*** MVenesio has joined #openstack-cinder13:07
*** sbezverk has joined #openstack-cinder13:09
TheJulia_heena_py_: I think it it might help if you elaborate context, because attachment is slightly overloaded as it is used in several places so your question could mean a couple different things.13:11
*** dikonoo has joined #openstack-cinder13:11
*** yangyapeng has quit IRC13:11
*** MVenesio has quit IRC13:13
*** MVenesio has joined #openstack-cinder13:13
*** josecastroleon has joined #openstack-cinder13:15
*** dikonoor has quit IRC13:15
*** randomhack has joined #openstack-cinder13:18
*** savvas has quit IRC13:21
*** bhagyashris has quit IRC13:22
*** savvas has joined #openstack-cinder13:22
*** randomhack has quit IRC13:23
*** randomhack has joined #openstack-cinder13:24
openstackgerritMerged openstack/cinder master: Fix wrong links
*** Lee1092 has quit IRC13:28
*** takedakn has joined #openstack-cinder13:46
*** jmlowe_ has joined #openstack-cinder13:47
*** jmlowe has quit IRC13:49
*** _pewp_ has quit IRC13:52
scottda_heena_py_: If you are using current HEAD of master, you can use V3 attachment APIs:
*** markvoelker_ has quit IRC13:52
scottdaIF you are not on current master, You have no API for attachments available.13:52
*** cknight has joined #openstack-cinder13:54
*** lucasxu has joined #openstack-cinder13:54
*** felipemonteiro has joined #openstack-cinder13:55
*** _pewp_ has joined #openstack-cinder13:55
jgriffithstvnoyes ildikov smcginnis morning update :)13:58
jgriffithdigging through things again last night the problem is actually pretty simple... and it's my own fault for being naive :)13:58
ildikovjgriffith: so what is it exactly?13:59
jgriffiththere is some potential to fix up the connection info (attachment_update) and that's worth doing... but the root cause of the various failures is a race between connect and detach13:59
ildikovjgriffith: pretty simple is of course music to my ears :)13:59
jgriffithildikov so the problem is that since we immediately toggle the volume to in-use when cinder gets the attachment-update call; all of our tests and every make assumptions based on that14:00
*** randomhack has quit IRC14:00
ildikovjgriffith: not every test is building on that info as I saw14:00
jgriffithso what we're seeing with things like the hanging devices, is because a detach is issues WHILE os-brick is still actually setting up the connection14:00
*** alkhodos has joined #openstack-cinder14:01
ildikovjgriffith: although I guess if we fix connection info we will not see the iscsi issues and not losing performance on it14:01
jgriffithildikov 'not every test is building on that info as I saw' ?  I'm unsure what that means?14:01
ildikovthat no matter what Cinder sets for the volume if no one checks whether or not it is 'in-use' before cleaning up14:01
jgriffithildikov oh... well not really though14:02
jgriffithbecause the tests pretty much always wait for appropriate status changes to do "the next thing"14:02
jgriffithso there's something else pretty dangerous about the way I have it right now with the auto-set thing....14:02
ildikovso I haven't checked the old flow, maybe I should but I would guess brick is not struggling that much with creating the actual connection14:03
jgriffithall of our tests assume that when that status is toggled that means nova is done and it succesfully made the connection14:03
jgriffithildikov I think the timing issues with brick and the retries may have been related to the connection_info data, but regardless it's something we should deal with anyway for busy systems etc14:03
jgriffithwhat I was going to mention about the "dangerous" thing is it's quite possible for a full test suite to run and everything pass without ever making a single iscsi connection14:04
jgriffithwell, not quite "every test" but a good portion of them14:04
jgriffithso that's sort of crappy...14:05
jgriffithor good depending on your point of view :)14:05
jgriffithif you just want tests to pass it's fantastic!! :)14:05
jgriffithso my proposal, and what I think we should do....14:05
ildikovno, my main concern is that I don't want this to be slower than the old flow was as that's a problem14:06
*** mvk_ has joined #openstack-cinder14:06
ildikovand yeah, attach_volume in tempest looks for 'in-use'14:06
jgriffithildikov oh.. it won't be after this is dealt with14:06
jgriffithat least I don't think it will be14:06
ildikovbut still, my bigger concern is the connection_info data and all those weird iscsiadm warnings/errors14:06
jgriffithso as much as I like the consolidated 2 step attach thing, we are going to need a 3'rd thing14:06
jgriffithildikov those are gone for me now14:07
ildikovso we will need a third call from Nova that the connection is done14:07
jgriffithyeah, probably so....14:07
ildikovwhich I guess means a version bump14:07
*** yangyapeng has joined #openstack-cinder14:08
*** mvk has quit IRC14:08
*** markstur has joined #openstack-cinder14:08
jgriffithwhich also means I might as well fix the responses while we're at it14:08
jgriffithThe only other option is to add internal tracking of that in Nova14:08
ildikovinternal tracking?14:08
jgriffithsomething like a status column in the bdm table14:08
jgriffithbdm.attach_status = detached|attaching|attached14:09
ildikovI would call it connection_status14:09
jgriffiththat way when nova gets a detach request, it looks at that and decides if it should wait, continue etc14:09
ildikovotherwise it's a hard sell why we need attach_status on both sides14:09
jgriffithYeah, I'm not particular about the name14:09
ildikovif we want to decouple Cinder a bit, then it's not a terribly bad option14:10
ildikoveven if I really hate touching the BDM14:10
jgriffithildikov yeah, it has pros and cons14:10
ildikovwhich one are you leaning towards?14:11
jgriffithmriedem may have some thoughts about that which could make the decision easy (like "no way")14:11
ildikovjgriffith: true :)14:11
jgriffithwell in terms of not rocking the boat and not introducing new problems, I'm inclined to add the third call to cinder14:11
jgriffithor add it as a form of attachment_update14:12
*** yangyapeng has quit IRC14:12
jgriffithso one thing I thought of was that attachment_update would take a flag and set to "attaching" instead of "in-use" if that flag was set14:12
*** markstur has quit IRC14:12
ildikovI would like to avoid flags if possible14:12
jgriffiththen attachment_update (or a new call altogether) could also be extended to be used for notification of connection complete14:13
jgriffithmeh, flags/args are ok if theyr'e not overused and they make sense14:13
jgriffithIMHO at least14:13
jgriffithbut yeah, I'm open14:13
ildikovyeah, they all start as one simple extra function and then turn into monsters :)14:14
jgriffithLOL.. fair enough14:14
ildikovanyway, maybe I just overcomplicate it :)14:17
*** markstur has joined #openstack-cinder14:20
jgriffithIt's funny, we rat holed on the whole "multiple calls to brick for a shared connection" thing so much I kinda missed the most obvious and blatant thing of all14:20
ildikovwell, I think we talked about it at a certain point a half year ago or so, but then let it go14:22
jgriffithprobably.. this had gone on far too long :)14:22
jgriffithok... well maybe I'll code up the changes and see how things go14:22
ildikovsounds good, I'll test it when you got something14:23
jgriffithok, sounds good14:23
*** psachin has quit IRC14:24
*** markstur has quit IRC14:24
ildikovin the meantime hopefully mriedem let's us know his preference on where to track things14:26
*** mdavidson has quit IRC14:26
jgriffithildikov yeah, I think I know the answer... and I think it's better to reflect it in cinder but we'll see14:27
*** armax has joined #openstack-cinder14:31
*** takedakn has quit IRC14:32
*** itlinux has joined #openstack-cinder14:34
openstackgerritLucian Petrut proposed openstack/cinder master: SMBFS: enable thin provisioning support flag
*** mdavidson has joined #openstack-cinder14:35
*** Julien-zte has quit IRC14:36
*** yangyapeng has joined #openstack-cinder14:44
mriedemi haven't read scrollback, but14:49
mriedem""bdm.attach_status = detached|attaching|attached"14:50
mriedemisn't that the completely wrong way to be going?14:50
mriedemi thought that's why we removed the check_attach and check_detach stuff,14:50
mriedemto stop coupling nova state tracking with cinder internal state14:50
mriedemwe need another call now on detach?14:52
mriedemor on attach?14:52
mriedemseems we're going back to the existing v2 APIs :) reserve, initialize_connection, attach14:52
jgriffithmriedem yep, pretty much14:52
mriedemnow we'll have attachment_create (reserve), attachment_update (initailize_connection), super-final-done-with-connect (attach)14:52
jgriffithmriedem si14:53
mriedemwhat does that buy us over what we had before?14:53
jgriffithmriedem just it's own API and the attachment-id14:53
jgriffithwhich it occured to me last night we likely could've done anyway14:53
mriedembut, does this still make multi-attach support more possible?14:54
*** crushil has joined #openstack-cinder14:54
jgriffithmriedem absolutely, yes... the attachment object is what makes that possible14:54
mriedemin the new apis rather than bolting that onto the old apis?14:54
mriedemjust need to keep reminding me of the justification here :)14:54
jgriffiththe old api is kind of a mess at this point, so that would be no good14:54
*** sbezverk has quit IRC14:54
jgriffithmriedem yeah... it's unfortunate; but there is still value14:55
*** felipemonteiro_ has joined #openstack-cinder14:55
jgriffithmriedem it is more simple, and it does allow some things to be more robust in the future14:55
*** felipemonteiro has quit IRC14:55
jgriffithmost of all, no more guessing about whether we can or can't disconnect something14:55
*** sbezverk has joined #openstack-cinder14:55
jgriffithwell... except this particular little race condition :)14:56
jgriffithI'd like to figure out a different way to solve it; but I don't want to rewrite all of our tests etc to do so14:56
*** rcernin has quit IRC14:57
jgriffithmriedem the only other thing is if we put a lock in nova while it's doing the os-brick connect, and that keeps the detach from coming in while that's in process?14:57
jgriffithbut I think that would be troublesome14:57
*** thingee_ has joined #openstack-cinder14:58
jgriffithsmcginnis jungleboyj then there's the whole snapshot-attach.. GRRRRRR14:59
* jungleboyj needs to read the backlog for context apparently.15:00
*** dikonoo has quit IRC15:02
jungleboyjjgriffith:  Is snapshot attach more complicated?15:03
jgriffithjungleboyj nah.... but this:
*** markstur has joined #openstack-cinder15:04
jgriffithjungleboyj we probably should've forced usage of the new API or waited IMO15:04
jgriffithjungleboyj now it's one more thing to go back and change15:04
jgriffithnot a huge deal, especially since this is really just an internal backup thing15:05
mriedemjgriffith: the only way we can do a lock like that in nova is via db15:05
mriedemif it's a shared connection i mean15:05
mriedemnot really sure though, johnthetubaguy and mdbooth went over this quite a bit when writing the spec15:05
mriedembut i didn't think a process lock on nova-compute would be sufficient15:06
jgriffithmriedem yeah... which goes back to the bdm entry again.. and honestly that isn't a great answer anyway.  Yeah... johnthetubaguy was definitely thinking about it a lot15:06
jgriffithmriedem and yes, I think you're correct; it would help, but I don't think it's good enough15:06
jgriffithand provides a great opportunity for "stuck in xxxxx" states15:06
mriedemi don't want to mention etcd either, even though i just did15:06
jgriffithwhich nobody wants to go back to15:07
jgriffithmriedem I'm ignoring that line15:07
jgriffithso I don't like the extra cinder call but honestly I think it's the right answer.15:08
jgriffithI'm coding it up now and then will get a bunch of testing on it and make sure there's not yet another problem15:08
jgriffithif there is... well then I'll just drink heavily15:09
*** savvas has quit IRC15:13
*** baojg_ has joined #openstack-cinder15:15
*** Dinesh_Bhor has quit IRC15:15
*** dklyle has joined #openstack-cinder15:15
*** david-lyle has quit IRC15:15
*** baojg has quit IRC15:16
openstackgerritMerged openstack/python-cinderclient master: Fix wrong links
jungleboyjIt will be good when we have this whole API cleanup done.  One step at a time though.15:22
jungleboyjjgriffith: I will need to spend some time at the PTG with you better understanding what has been fixed and where we are at.15:23
jgriffithjungleboyj for sure, no problem15:23
*** mvk has joined #openstack-cinder15:24
jgriffithjungleboyj depending on time, might even be worth reviewing the whole thing as a team15:24
jungleboyjjgriffith: Cool, don't want to be a pest.  Ildiko has also been doing such a good job of leading I haven't worried much.15:24
jungleboyjjgriffith: Ok, I didn't want to ask for that if everyone felt up to date, but doing a whole team review would be awesome.15:25
jungleboyjI can put it on the agenda and see what everyone thinks.15:25
jgriffithyeah, ildikov is our fearless leader on this one :)15:25
*** mvk_ has quit IRC15:25
*** e0ne has quit IRC15:27
*** dikonoo has joined #openstack-cinder15:29
*** mvk has quit IRC15:29
stvnoyesjgriffith,mriedem: since the overlapping of detach too quickly following an attach is a problem, is there a way for nova to know an attach's connect is not complete and throw out of a premature detach (and vice versa)? The client can do that if they use events. but we can't rely on the client doing that.15:29
*** mvk has joined #openstack-cinder15:29
jgriffithstvnoyes I looked briefly at that, but couldn't figure out how to do it without state changes in nova-db; but admittedly there could be another way15:31
jgriffithstvnoyes I just put a delay in nova/compute/api:detach for 5 seconds... that works too :)15:31
stvnoyesyeah, I found that a delay in the test worked too; very clean ;-) (not)15:32
jgriffithyeah, I stole your idea and just made it global :)15:32
jgriffithconsidered hiding it in a patch :)15:32
stvnoyesyou're only supposed to steal my good ideas!15:33
jgriffithstvnoyes I don't know... it was looking like a pretty good idea at the time :)15:33
*** dikonoo has quit IRC15:33
*** ociuhandu has joined #openstack-cinder15:34
*** wN has joined #openstack-cinder15:41
*** wN has quit IRC15:41
*** wN has joined #openstack-cinder15:41
*** dikonoo has joined #openstack-cinder15:42
*** itlinux has quit IRC15:46
*** aarefiev is now known as aarefiev_afk15:47
*** savvas has joined #openstack-cinder15:48
*** Apoorva has joined #openstack-cinder15:58
*** dikonoo has quit IRC15:59
*** ntpttr_laptop has joined #openstack-cinder16:01
*** armax has quit IRC16:04
*** sbezverk has quit IRC16:05
*** lucasxu has quit IRC16:07
jungleboyjGotta love software held together with sleeps.  :-)16:07
*** ociuhandu has quit IRC16:10
*** dikonoo has joined #openstack-cinder16:17
*** lpetrut has quit IRC16:18
*** markvoelker has joined #openstack-cinder16:25
*** itlinux has joined #openstack-cinder16:33
*** jmlowe_ has quit IRC16:34
ildikovjgriffith: mriedem: sorry, was afk a bit16:36
ildikovjgriffith: mriedem: while I'm not happy about the new call, sounds like the right thing to do16:36
*** psachin has joined #openstack-cinder16:36
ildikovjungleboyj: I think a team review would be very useful as we need to figure out what multi-attach brings into the picture too and it would be great having people aware of the base API before we dig deeper into that16:38
*** savvas_ has joined #openstack-cinder16:39
*** tesseract has quit IRC16:41
*** savvas has quit IRC16:41
jungleboyjildikov:  Good.  Glad you like that plan.  I have added it to the Agenda.16:41
ildikovjungleboyj: great, thank you16:42
ildikovjungleboyj: how packed the agenda is?16:42
jungleboyjildikov:  Not particularly packed right now.  There are a few items that will take longer discussion, but we definitely have room for more topics.16:43
ildikovjungleboyj: I might add an optional multi-attach item hoping that we can sort out the remaining issues with the new API rather quickly16:43
jungleboyjildikov:  Great idea!16:43
ildikovjungleboyj: sounds good16:43
ildikovjungleboyj: do you have the etherpad link handy?16:43
jungleboyjildikov: Yep.16:43
ildikovjungleboyj: thank you16:44
*** dklyle has quit IRC16:44
*** dklyle has joined #openstack-cinder16:45
*** jmlowe has joined #openstack-cinder16:56
*** markvoelker has quit IRC16:59
*** Apoorva_ has joined #openstack-cinder17:01
*** e0ne has joined #openstack-cinder17:02
*** Apoorva has quit IRC17:04
*** e0ne has quit IRC17:05
*** e0ne has joined #openstack-cinder17:07
*** psachin has quit IRC17:08
*** e0ne_ has joined #openstack-cinder17:09
*** Swanson has quit IRC17:12
*** e0ne has quit IRC17:12
*** abhitechie has joined #openstack-cinder17:13
*** abhitechie has quit IRC17:14
*** felipemonteiro_ has quit IRC17:14
*** abhitechie has joined #openstack-cinder17:15
*** mvk has quit IRC17:17
*** bradbeam has joined #openstack-cinder17:17
*** bradbeam has left #openstack-cinder17:20
*** dikonoor has joined #openstack-cinder17:21
*** diablo_rojo has joined #openstack-cinder17:22
*** ntpttr_laptop has quit IRC17:23
*** dikonoo has quit IRC17:24
*** vivsoni_ has quit IRC17:28
*** savvas_ has quit IRC17:29
*** savvas has joined #openstack-cinder17:29
*** abhitechie has quit IRC17:32
*** abhitechie has joined #openstack-cinder17:33
*** simondodsley has joined #openstack-cinder17:33
simondodsleyAnyone got decent documentation on 'cinder manage'? There seems to be a lack of good info out there. Wanted to check before I started delving into the code to work out how to use it correctly.17:34
jgriffithstvnoyes ildikov prelim patches here:
jgriffithwould love any feedback if you see something bad in there17:35
*** e0ne_ has quit IRC17:35
jgriffithsimondodsley god be with you17:36
*** abhitechie has quit IRC17:37
*** abhitechie has joined #openstack-cinder17:37
simondodsleyjgriffith I know - looking for the easy option first :)17:37
patrickeastsimondodsley: what? that obscure jira reference and suggestion to use "cinder help manage" wasn't enough? :P17:38
jgriffithsimondodsley I think this about as good as it gets:
jgriffithbut I'm sure you already saw that17:39
patrickeasthah yea17:39
*** crushil has quit IRC17:39
jgriffithwrong cinder-manage!!!!17:39
simondodsleygoing to say...17:39
patrickeastthats why i seriously regret the naming, its impossible to search for "cinder manage" help17:39
patrickeastiirc it did get a new name with osc too... although i don't know what it is17:40
simondodsley@patrickeast: not the most helpful JIRA17:40
*** abhitechie has quit IRC17:40
*** abhitechie has joined #openstack-cinder17:40
patrickeastsimondodsley: theres a comment about halfway down where i show how to use it17:41
jgriffithof course at this point I'm resorting to google so I have no value add17:41
jgriffithavishay is your target17:41
simondodsleywould be nice if the actual openstack docs talked about this, not something from NetApp17:41
jgriffithsimondodsley I don't disagree.... but hey... we ARE OpenStack :)17:41
patrickeastthere is one in the admin guide about managing snapshots17:41
patrickeastpretty close17:41
*** e0ne has joined #openstack-cinder17:42
simondodsleypatrickeast: at least that's vaguely correct. The docs we have on our support site reference the old Kilo syntax... I had to remove them this morning17:42
patrickeastyea, theres a bug open for that somewhere...17:43
ildikovjgriffith: looks good overall17:43
jgriffithildikov amazingly, so far it even seems to work :)17:43
ildikovjgriffith: even better :)17:43
*** e0ne has quit IRC17:43
jgriffithhope I didn't just jinx it17:44
patrickeasti don't know how to update them though... we should just move it to
*** ntpttr_laptop has joined #openstack-cinder17:44
patrickeastsimondodsley: ^^17:44
jgriffithso basically we did a whole lot of work to reimplement the old attach flow :)17:44
ildikovjgriffith: by seeing the console output I thought you're kinda certain :)17:44
jgriffithildikov that was my manual testing; now it's the moment of truth17:45
ildikovjgriffith: tempest, yaaay :)17:46
*** abhinavtechie has joined #openstack-cinder17:51
*** abhitechie has quit IRC17:51
*** crose has joined #openstack-cinder17:53
*** markvoelker has joined #openstack-cinder17:55
*** Apoorva_ has quit IRC18:03
*** crushil has joined #openstack-cinder18:03
*** Apoorva has joined #openstack-cinder18:03
*** thingee_ has quit IRC18:07
*** e0ne has joined #openstack-cinder18:12
*** e0ne has quit IRC18:12
*** sdague has quit IRC18:17
*** sdague has joined #openstack-cinder18:17
*** savvas has quit IRC18:19
*** dobson has quit IRC18:23
*** wN_ has quit IRC18:23
*** dobson has joined #openstack-cinder18:26
ildikovjgriffith: smcginnis: jungleboyj: we should get the cinderclient changes for what we agreed above in too before we cut the next release18:27
*** markvoelker has quit IRC18:29
*** wN_ has joined #openstack-cinder18:29
*** abhinavtechie has quit IRC18:29
*** abhitechie has joined #openstack-cinder18:29
*** savvas has joined #openstack-cinder18:31
*** savvas has quit IRC18:31
*** savvas has joined #openstack-cinder18:32
*** diablo_rojo has quit IRC18:33
*** diablo_rojo has joined #openstack-cinder18:33
jungleboyjIldikov sounds good.18:38
jgriffithildikov ok... we look pretty good now.  All tests passing, no performance hit and clean logs!18:40
ildikovjgriffith: sounds awesome!!!18:40
ildikovjgriffith: thank you!18:40
jgriffithguess I'll start making official versions of the cinder patches..... only remaining question is "to translate, or not-to translate" :)18:40
ildikovjgriffith: that's a tough one18:41
*** nicolasbock has quit IRC18:43
*** abhitechie has quit IRC18:45
ildikovjgriffith: I kinda lean towards switching back to the original version18:45
*** crose has quit IRC18:45
jgriffithildikov it's certainly easier18:46
jgriffithok... I'll try that18:46
ildikovjgriffith: but if we're certain that the old structure doesn't have a purpose what we have now is a good start to remove it18:46
*** crushil has quit IRC19:04
diablo_rojohemna, jungleboyj I am writing an article for superuser about how people can follow cinder's ongoings at the ptg so other projects can read it and do similar things. Any interest in helping/reading over a draft early next week?19:07
jungleboyjdiablo_rojo:  Sure, I can do that.19:07
diablo_rojojungleboyj, cool. I will have a draft for you on monday to add to/make comments on etc.19:07
diablo_rojoI want to have it done by Wednesdayish so it can get posted before the end of the week.19:08
jungleboyjdiablo_rojo:  Sounds good.19:08
jungleboyjdiablo_rojo:  Ok, shouldn't be a problem.19:08
diablo_rojojungleboyj, cool :) Good way to get Cinder in the limelight a bit and teach other projects how we do things virtually for those that can't attend.19:09
jungleboyjdiablo_rojo:  Oh yeah.  That would be good for us to share.  Should talk about how we are trying to make our notes and action items more actionable after the PTG as well.19:11
jungleboyjdiablo_rojo:  Get me a draft and I will share my thoughts.19:11
diablo_rojoGood idea. I'm just working on an intro and outline right now and then I can share it with you.19:11
diablo_rojojungleboyj ^^19:12
*** crushil has joined #openstack-cinder19:12
jungleboyjdiablo_rojo:  Cool.19:12
*** ntpttr_laptop has quit IRC19:18
*** geguileo is now known as geguileo_PTO19:19
*** ChubYann has joined #openstack-cinder19:23
*** lpetrut has joined #openstack-cinder19:25
*** markvoelker has joined #openstack-cinder19:27
*** Apoorva has quit IRC19:32
*** adisky__ has quit IRC19:39
*** ntpttr_laptop has joined #openstack-cinder19:45
diablo_rojojungleboyj, I shared it with you.19:49
jungleboyjdiablo_rojo: Thanks.  I see the request.  Will take a look soon.19:52
diablo_rojojungleboyj, its a bit more than an outline :) I got writing and couldn't stop.19:55
jungleboyjdiablo_rojo:  Looks like a good start.  I will look when I don't have so much going on.19:58
*** markvoelker has quit IRC19:59
diablo_rojojungleboyj, thanks, sounds good.20:00
*** wN_ has quit IRC20:01
*** lpetrut has quit IRC20:01
*** lpetrut has joined #openstack-cinder20:02
*** wN_ has joined #openstack-cinder20:04
*** gouthamr has quit IRC20:14
*** MVenesio has quit IRC20:14
*** catintheroof has quit IRC20:15
*** Apoorva has joined #openstack-cinder20:16
*** gouthamr has joined #openstack-cinder20:33
*** jmlowe has quit IRC20:45
*** itlinux has quit IRC20:46
*** jmlowe has joined #openstack-cinder20:47
*** dikonoor has quit IRC20:51
*** cknight has quit IRC20:53
*** chhavi has quit IRC20:53
*** jmlowe has quit IRC20:56
*** markvoelker has joined #openstack-cinder20:56
*** abishop has quit IRC21:00
*** Apoorva_ has joined #openstack-cinder21:00
*** dustins has quit IRC21:00
*** edmondsw has quit IRC21:01
*** jistr is now known as jistr|off21:02
*** Apoorva has quit IRC21:04
*** crushil has quit IRC21:06
*** dave-mccowan has quit IRC21:12
*** crushil has joined #openstack-cinder21:14
*** markvoelker has quit IRC21:30
*** MVenesio has joined #openstack-cinder21:33
*** baojg has joined #openstack-cinder21:37
*** baojg_ has quit IRC21:37
*** Apoorva_ has quit IRC21:38
*** Apoorva has joined #openstack-cinder21:38
*** ntpttr_laptop has quit IRC21:40
*** xyang1 has quit IRC21:54
*** markvoelker has joined #openstack-cinder22:10
*** ntpttr_laptop has joined #openstack-cinder22:16
*** lbragstad has quit IRC22:17
*** MVenesio_ has joined #openstack-cinder22:18
*** MVenesio has quit IRC22:22
*** lpetrut has quit IRC22:23
*** markvoelker has quit IRC22:33
*** MVenesio_ has quit IRC22:51
openstackgerritChaynika Saikia proposed openstack/cinder master: cinder incremental backup with ceph fails if last one deleted
*** sdague has quit IRC23:09
smcginnisjgriffith: Finally home and online. Reading scrollback. Not loving another call, but that's great if it's all sorted out and working now.23:17
smcginnissimondodsley: OSC renamed manage to "adopt" and "abandon" IIRC.23:17
openstackgerritJohn Griffith proposed openstack/cinder master: Make attachment_update set status to attaching not
smcginnisjgriffith: Is "attaching not" a new status for not attached? :D23:36
jgriffithsmcginnis LOL.. you got me23:39
jgriffithI suppose I should fix that :)23:40
smcginnisEh, maybe wait and see if you need to do another update?23:40
openstackgerritJohn Griffith proposed openstack/cinder master: Make attachment_update set status to attaching
jgriffithsmcginnis already done... and no more updates!!!23:41
smcginnisjgriffith: +1 let's hope!23:41
jgriffithclient on the way23:41
jgriffithwe can talk Monday about backport, etc :(23:41
jgriffithI don't know how that works with api micro-v's23:41
jgriffithif at all23:41
openstackgerritJohn Griffith proposed openstack/python-cinderclient master: Add an attachment_complete API call
smcginnisYeah... not sure if that's really an option at this point. :/23:42
jgriffithOk... fair enough23:42
smcginnisMaybe in RC2.23:42
jgriffithon that note, I gotta run; I'll catch ya later.23:42
smcginnisHave a good weekend.23:42
jgriffithThanks!!!  You too!!23:42
*** zzzeek has quit IRC23:43
*** zzzeek has joined #openstack-cinder23:44
*** Apoorva_ has joined #openstack-cinder23:50
*** Apoorva has quit IRC23:54
*** Apoorva_ has quit IRC23:55

Generated by 2.15.3 by Marius Gedminas - find it at!