Tuesday, 2016-12-20

« Monday, 2016-12-19 Index Wednesday, 2016-12-21 »
*** catintheroof has quit IRC00:29
*** chlong has quit IRC00:59
openstackgerritPony Chou proposed openstack/manila: Add QNAP Manila Driver  https://review.openstack.org/39470301:38
*** tuanluong has joined #openstack-manila01:46
openstackgerritzhongjun proposed openstack/manila: Trivial fix LOG.exception issues  https://review.openstack.org/41268801:56
*** gouthamr has quit IRC02:10
*** gcb has quit IRC02:30
openstackgerritHa Van Tu proposed openstack/manila: Fix typo in rootwrap.conf  https://review.openstack.org/41270202:31
*** gcb has joined #openstack-manila02:50
openstackgerritgengchc2 proposed openstack/manila: use six.StringIO for compatibility with io.StringIO in python3  https://review.openstack.org/41272303:21
openstackgerritzhongjun proposed openstack/manila: Enable IPv6 in manila(network plugins and drivers)  https://review.openstack.org/40677603:23
openstackgerritzhongjun proposed openstack/manila: Add the ability to check the tenant quota in detail  https://review.openstack.org/36661403:48
*** senk has joined #openstack-manila05:27
*** yankee has joined #openstack-manila05:39
*** yankee has quit IRC05:40
openstackgerritPony Chou proposed openstack/manila: Add QNAP Manila Driver  https://review.openstack.org/39470305:52
*** senk has quit IRC05:59
*** lpetrut has joined #openstack-manila06:37
*** dsariel has joined #openstack-manila06:38
*** xinyanzhang has joined #openstack-manila06:43
*** dsariel has quit IRC07:30
*** lpetrut has quit IRC07:32
*** nherciu has joined #openstack-manila08:06
openstackgerritPony Chou proposed openstack/manila: Add QNAP Manila Driver  https://review.openstack.org/39470308:14
*** rraja has joined #openstack-manila08:16
*** pcaruana has joined #openstack-manila08:19
*** dsariel has joined #openstack-manila08:51
openstackgerrityunfeng zhou proposed openstack/puppet-manila: add CONTRIBUTING.rst  https://review.openstack.org/41291808:58
*** lpetrut has joined #openstack-manila09:03
*** lpetrut has quit IRC09:09
*** lpetrut has joined #openstack-manila09:09
*** a-pugachev has joined #openstack-manila09:15
tommylikehuping vponomaryov09:17
*** lpetrut1 has joined #openstack-manila09:43
*** lpetrut has quit IRC09:43
*** lpetrut1 is now known as lpetrut09:43
*** lpetrut has quit IRC09:48
*** lpetrut has joined #openstack-manila09:56
*** ganso has joined #openstack-manila09:57
openstackgerritHa Van Tu proposed openstack/manila: Fix typo in rootwrap.conf  https://review.openstack.org/41270210:07
*** ociuhandu has quit IRC10:17
*** surabujin has joined #openstack-manila10:19
*** senk has joined #openstack-manila10:35
vponomaryovtommylikehu: pong10:37
*** gcb has quit IRC10:59
*** senk has quit IRC11:00
openstackgerritMerged openstack/manila: GPFS CES: Fix bugs related to access rules not found  https://review.openstack.org/41101011:00
openstackgerritMerged openstack/manila: [TrivialFix] Move share type filter tempest to test_scheduler_stats.py  https://review.openstack.org/40964111:01
*** igajsin has joined #openstack-manila11:01
*** dsariel has quit IRC11:07
*** igajsin has quit IRC11:17
*** alyson_ has joined #openstack-manila11:21
openstackgerritTom Barron proposed openstack/manila: WIP: Remove nova net from service_instance  https://review.openstack.org/41044911:23
*** akapil has joined #openstack-manila11:27
*** igajsin has joined #openstack-manila11:33
*** akapil has quit IRC11:37
*** akapil has joined #openstack-manila11:38
*** ociuhandu has joined #openstack-manila11:53
*** tuanluong has quit IRC11:58
*** akapil has quit IRC12:04
*** akapil has joined #openstack-manila12:04
*** senk has joined #openstack-manila12:06
*** senk has quit IRC12:08
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] debug 3.0  https://review.openstack.org/41036612:12
*** catintheroof has joined #openstack-manila12:37
*** ociuhandu has quit IRC12:43
*** ociuhandu has joined #openstack-manila12:43
*** lseki has joined #openstack-manila12:43
*** dsariel has joined #openstack-manila12:47
*** ociuhandu has quit IRC12:48
openstackgerritValeriy Ponomaryov proposed openstack/python-manilaclient: [DNM] Test share group client changes  https://review.openstack.org/41258712:49
*** chlong has joined #openstack-manila12:53
*** ociuhandu has joined #openstack-manila12:54
*** gouthamr has joined #openstack-manila12:55
*** ociuhandu has quit IRC12:59
*** pcaruana has quit IRC13:01
*** pcaruana has joined #openstack-manila13:06
*** ociuhandu has joined #openstack-manila13:09
*** ociuhandu has quit IRC13:14
*** ociuhandu has joined #openstack-manila13:24
*** akapil has quit IRC13:32
*** ociuhandu has quit IRC13:38
*** ociuhandu has joined #openstack-manila13:38
*** akapil has joined #openstack-manila13:44
openstackgerritMerged openstack/manila: Fix typo in rootwrap.conf  https://review.openstack.org/41270213:49
*** ociuhandu has quit IRC13:50
*** catinthe_ has joined #openstack-manila14:06
*** mtanino has quit IRC14:08
*** senk has joined #openstack-manila14:08
*** catintheroof has quit IRC14:08
*** chlong has quit IRC14:16
*** surabujin has quit IRC14:20
*** cknight has joined #openstack-manila14:22
*** pcaruana has quit IRC14:23
*** cknight1 has joined #openstack-manila14:25
*** cknight has quit IRC14:26
*** surabujin has joined #openstack-manila14:33
*** dustins has joined #openstack-manila14:35
*** xyang_ has joined #openstack-manila14:36
*** pcaruana has joined #openstack-manila14:37
*** senk has quit IRC14:37
*** eharney has quit IRC14:43
*** ociuhandu has joined #openstack-manila14:47
openstackgerritMerged openstack/manila: use six.StringIO for compatibility with io.StringIO in python3  https://review.openstack.org/41272314:50
*** cknight1 has left #openstack-manila15:03
*** eharney has joined #openstack-manila15:04
*** eharney has quit IRC15:12
dustinsbswartz: ping15:13
dustinsabout PTG stuff15:13
*** jcsp has quit IRC15:19
*** jcsp has joined #openstack-manila15:19
openstackgerritzhongjun proposed openstack/manila: Trivial fix LOG.exception issues  https://review.openstack.org/41268815:23
*** dustins_ has joined #openstack-manila15:30
*** dustins has quit IRC15:32
*** mtanino has joined #openstack-manila15:32
*** eharney has joined #openstack-manila15:34
*** dustins_ is now known as dustins15:34
bswartzdustins: yes?15:35
*** xyang_ has quit IRC15:35
dustinsbswartz: Heya, what do you think the over/under is for the PTG meetings going into Friday?15:35
bswartzdustins: we've already said manila will be wed/thurs only15:36
tommylikehu_bswartz: could you take some time on this patch: https://review.openstack.org/#/c/406776/15:36
dustinsbswartz: So Thursday is the hard stop?15:36
dustinsNo chance of going into Friday?15:36
bswartzdustins: I personally plan to jump over to cinder on Friday15:36
bswartztommylikehu_: you've got some -1s15:37
dustinsbswartz: Okay, cool, just trying to see when I need to head out of Atlanta and wanted to see if there was a possibility of it going long15:37
bswartztommylikehu_: anything specific you want me to focus on?15:37
dustinsbswartz: Thanks!15:37
tommylikehu_yes and I also want some feedback from you with which I can fix them all next morning :)15:39
*** xyang_ has joined #openstack-manila15:39
*** akapil has quit IRC15:41
*** a-pugachev has quit IRC15:42
*** xyang_ has quit IRC15:43
*** akapil has joined #openstack-manila15:45
*** senk has joined #openstack-manila15:46
*** harlowja has joined #openstack-manila15:49
*** lseki has quit IRC15:50
*** chlong has joined #openstack-manila15:53
*** xyang_ has joined #openstack-manila15:55
*** tommylikehu_ has quit IRC15:56
*** akapil has quit IRC15:58
*** rraja is now known as rraja|afk15:59
*** chlong has quit IRC16:01
openstackgerritVitaliy Levitski proposed openstack/manila: Add MapR-FS native driver  https://review.openstack.org/35003516:01
*** akapil has joined #openstack-manila16:02
*** xyang_ has quit IRC16:03
*** xyang_ has joined #openstack-manila16:11
*** chlong has joined #openstack-manila16:17
*** tommylikehu_ has joined #openstack-manila16:24
*** senk has quit IRC16:26
*** chlong has quit IRC16:28
*** xyang_ has quit IRC16:29
*** xyang_ has joined #openstack-manila16:34
*** catintheroof has joined #openstack-manila16:39
*** catinthe_ has quit IRC16:42
*** harlowja has quit IRC16:42
*** chlong has joined #openstack-manila16:42
gansobswartz, gouthamr, vponomaryov, toabctl, tbarron, markstur: Hello folks. Could you please review https://review.openstack.org/#/c/406305/ when you have some review time? Thanks in advance16:51
bswartzganso: ack, but my review queue is getting pretty deep16:51
gansobswartz: ok16:55
*** akapil has quit IRC16:58
*** dustins_ has joined #openstack-manila17:16
*** eharney has quit IRC17:17
*** eharney has joined #openstack-manila17:18
*** lpetrut has quit IRC17:18
*** dustins has quit IRC17:19
*** porrua has quit IRC17:24
*** eharney has quit IRC17:24
*** eharney has joined #openstack-manila17:38
*** nherciu has quit IRC17:51
*** harlowja has joined #openstack-manila17:53
*** lpetrut has joined #openstack-manila17:54
*** pcaruana has quit IRC17:59
*** lpetrut has quit IRC18:00
*** catinthe_ has joined #openstack-manila18:03
*** catintheroof has quit IRC18:06
*** xyang_ has quit IRC18:07
*** nherciu has joined #openstack-manila18:09
*** lpetrut has joined #openstack-manila18:12
*** xyang_ has joined #openstack-manila18:16
gansobswartz, gouthamr: ping18:20
gouthamrganso: pong18:20
gouthamrbswartz18:20
bswartzpong18:20
bswartzI was just rereading https://review.openstack.org/#/c/321213/18:20
gansobswartz: I would like to talk about mountable snapshots18:20
bswartzhttp://specs.openstack.org/openstack/manila-specs/specs/ocata/mountable-snapshots.html18:21
*** xyang_ has quit IRC18:21
bswartzganso: gouthamr mentioned there is a problem with the access rules driver interface in that spec18:22
*** porrua has joined #openstack-manila18:22
bswartzit looks like we missed this in spec review (or at least I did)18:22
gansobswartz: it uses the old allow/deny approach18:22
gansobswartz: it does not use update_access18:22
gansobswartz: I talked to tpsilva back then18:23
*** xyang_ has joined #openstack-manila18:23
gansobswartz: his motivation was that, back then in Newton the update_access was filled with race conditions and problems18:23
gansobswartz: so he wanted to have it the old way18:23
gansobswartz: as it was simpler and worked before update_access was introduced18:24
bswartzganso: well if we did it the old way we'd have to update it eventually to match the new way18:24
openstackgerritVitaliy Levitski proposed openstack/manila: Add MapR-FS native driver  https://review.openstack.org/35003518:24
bswartzand if we did that we'd end up with a mess similar to what we have now with driver not switching to the new interface fast enough18:25
gansobswartz: since it is a driver interface, it would not touch the API, as we did not touch the API when transitioning share rules to update_access18:25
gansobswartz: yes, very painful mess18:25
gouthamrganso: update_access affected the way instances work... if you have instances created in a deferred manner, allow/deny does not scale.18:25
bswartzganso: yes the _most_ important thing is that the REST API remains stable18:25
bswartzhowever significant driver interface changes are also pretty bad18:25
bswartzand should be avoided18:25
bswartzgiven that, my personal feeling is that we should update the spec to match the newer access rule driver interface18:26
bswartzand I'm wondering if anyone feels that's the wrong thing to do18:26
gansobswartz: first, is there any motivation accept the old approach?18:27
gansobswartz: like, I would go for that, we if do not have plans to change to update_access later18:28
ganso<bswartz> ganso: well if we did it the old way we'd have to update it eventually to match the new way18:28
gansobswartz: why exactly we would have to update?18:28
bswartzganso: I don't like the old approach for all of the same reasons we changed the access rules interface to being with18:28
gouthamrganso: one case: assume you have a snapshot, and you have to create a new snapshot instance, how do you update rules on this new instance?18:29
bswartzit's a fundamentally fragile way for the manager to interact with the backend18:29
bswartzif all you can do is allow and deny, then there's no way to do error recovery18:29
bswartzor even to validate that the current access rules are correct18:29
gansogouthamr: you duplicate the ShareSnapshotInstanceAccessMappings and apply them18:29
gouthamrganso: yep.. applying them could be one step: "update_access" or a loop over "allow_access"18:30
gansogouthamr: yes18:30
bswartzalso, the allow/deny interface doesn't support pushing multiple changes atomically18:30
gansobswartz: this time we would only have read-only access rules18:30
gansobswartz: it is simpler18:30
gansobswartz: yes, we would not support that with the old approach, only one at a time18:30
gansobswartz: it would be very limited18:31
gansobswartz: my question is: do we want to stay limited? that was another of tpsilva's arguments18:31
*** xyang_ has quit IRC18:31
gansobswartz: I told him the same thing and he said he designed the feature for special cases like, the admin wants to add an access rule to retrieve a file or something18:32
bswartzganso: that's the use case, but the design can support a lot more than that -- and that's the problem18:32
gouthamrganso: this feature is tenant-facing...  atleast that's what the spec says..18:32
gansobswartz: the way he envisioned the feature was to not let the mountable_snapshot access rules be used like share access rules18:32
bswartzif we allow for multiple access rules to be added, and we allow for rules to be removed, then we need to ensure those command actually work18:33
gansobswartz: I don't agree with him, but we have a spec merged according to this design and we are too close to FF now18:33
bswartza different design would be to have exactly 1 access rule per snapshot18:33
bswartzthat would be simpler and cover the use case you mentioned above18:33
bswartzhowever I feel that would be too limiting18:34
gansobswartz: yes, so we should re-imagine what we want mountable snapshots to be18:35
gansobswartz: decide to live with a more limited feature and never allow it to expand because it will be alot of work18:35
bswartzwell the spec says end users have complete freedom to add delete unlimited rules18:35
bswartzI didn't think that would be too much work, honestly18:36
gansobswartz: or add a more sophisticated and complex mechanism that is bound to have the same problems as the share access rules (which will be sorted out when share access rules are sorted out as well)18:36
bswartzI though the code could share a lot with the share access control code18:36
gansobswartz: in that case we could change the spec to be more limiting, or to change to the new approach18:36
gansobswartz: I think it is hard to share code, at this moment18:37
gansobswartz: that would be possible if we let the access rules refactor sink in first18:37
bswartzis the reason we perceive it to be a lot of work because it depends on the unfinished access rules work?18:37
gansobswartz: yes, that as well18:37
gansobswartz: the mountable snapshots would have to come 1 release later18:37
bswartzokay so what if we decide to do that? was there urgency to get the mountable snapshots feature in?18:38
gouthamrganso: hmmm, don't think so.. how about implementing an "update_snapshot_access" method and RPC and then worrying about any improvements later?18:38
*** ociuhandu has quit IRC18:39
gansogouthamr: bswartz is suggesting that the share access rules and snapshot access rules should converge. I agree with him, because else we would have to maintain 2 different code branches that do similar things18:39
gouthamrganso: i mean wrt the additional state transitions and all that.. though i don't see why just applying the access rules design will not work over there from the beginning18:39
gouthamrconverge where? driver interface?18:40
bswartzyes ganso -- copy/paste would give us correct behavior perhaps but the maintenance would be ugly -- long term it's best to simply share code for share/snapshot access control18:40
gansogouthamr: at the access rule helper18:41
gansogouthamr: driver interface will also be similar, but not the same18:41
gansogouthamr: looking at your access rule refactor code, do you feel like creating a more generic version of it that supports snapshot access rules in ocata timeframe?18:42
gansobswartz, gouthamr: I think that's the correct way to go, but we don't have much time18:43
bswartzganso: that's why I asked about the urgency18:44
bswartzif we make it depend on gouthamr's access rules changes then we have to accept that it's likely to slip into pike18:44
bswartzeven assume gouthamr can get the access rules change done in ocata -- they might come too late18:44
gouthamryes. i don't want to hold up tpsilva on that - spread too thin to prioritize this work so far.18:45
gansobswartz: which day is the last day we will be accepting new code?18:46
gouthamrhowever, the design's already up there in the spec. i don't mind some code duplication for the time being18:46
bswartzR-4 is feature freeze, so R-6 will be feature proposal freeze18:46
gansobswartz: so I will have 1 week after returning from vacation to get it done18:47
gouthamrganso: the alternate proposal is to let this merge but have no drivers except LVM support it right away and merge a fix for this in Pike?18:50
gansobswartz: I will return from vacation Jan 4th18:50
gansobswartz: my vacation starts this Saturday18:50
bswartzganso: the problem goes away if we slip this feature to pike -- I'm trying to understand if that's a problem or not18:51
bswartzyes my vacation starts saturday too, and I'll get back Jan 3rd18:51
gansogouthamr: HNAS driver would be merging in as well, I would be going through the pain of re-implementing the interface changes in Pike18:51
gansobswartz: sorry I return from vacation Jan 5th, not 4th18:51
gansogouthamr: but that is ok18:52
gouthamrganso: yes, i suppose that's the pain to take for iterative development of this feature anyway..18:52
gansogouthamr: only LVM and HNAS18:52
gansogouthamr: if more drivers want to implement that in Ocata, that would be more complicated18:52
gansobswartz, gouthamr: ok let's summarize what we have decided18:54
gouthamrganso: i wish we can merge with LVM, HNAS is up to you... but we could discuss this in the weekly meeting and let them know it's not completely ready18:54
gouthamrganso: we're hoping to finish the impl in Pike18:54
gansobswartz, gouthamr: 1) we want it with update_access. 2) we can have duplicated code and fix/converge it later if it has problems, what is more important is the driver interface and API18:54
gouthamrbswartz would call this an "experimental feature" instead of an experimental API18:55
bswartzganso: yes I agree with 1 and 218:55
bswartzgouthamr: what?18:55
gansogouthamr: I think there is no such term "experimental feature" right now18:55
gouthamr^ i'm making that up18:55
gansogouthamr: once a feature is in with driver interfaces, it is open to vendors to implement18:55
gansogouthamr: there is no way to say "don't implement it, it is experimental"18:56
gansogouthamr: because there will already be at least 1 driver supporting18:56
*** chlong has quit IRC18:56
gouthamrganso: hmmm, agree. 1) and 2) are fine by me too'18:57
gansobswartz, gouthamr: changing the implementation in Pike would cause all drivers to depend on the driver interface change, which would drag things towards Pike's FF18:57
gansobswartz, gouthamr: so, we need at least driver interface and API using update_access now in Ocata18:58
gansobswartz, gouthamr: whatever is in the middle can be improved in Pike18:58
gansobswartz, gouthamr: scratch what I said about update_access in API, it is allow/deny in API18:58
gansobswartz, gouthamr: what I meant is that we need it consolidated18:59
bswartzganso: the other option is to not merge anything until pike, but your proposal works too18:59
gouthamr+118:59
gansobswartz, gouthamr: ok, let's do the following, I will code all I can until Jan 12th19:00
*** rraja|afk is now known as rraja19:00
gansobswartz, gouthamr: but I won't prioritize a spec update that should be agreed before coding, the changes will be in the code19:00
*** martyt has joined #openstack-manila19:00
*** martyt has quit IRC19:00
gansobswartz, gouthamr: if I cannot get it done in time, it will slip to Pike19:01
gansobswartz, gouthamr: and it will not depend on goutham's patch directly19:01
*** martyt has joined #openstack-manila19:02
gansobswartz, gouthamr: ok?19:02
gouthamrganso: yep, works19:02
bswartzganso: why not update the spec?19:02
bswartzyou're concerned it will take too long, or you're concerned more changes will come?19:02
gansobswartz: I will update the spec, but not a requirement to be done before FF19:02
gansobswartz: maybe after Jan 12th19:02
bswartzpersonally I'd like to see the spec updated first, but I won't demand it19:03
bswartzI think the reviewers screwed up here (including me)19:03
gansobswartz: someone else in my team has just started working on manila-ui19:04
gansobswartz: I was going to start working on that, but now I will do this access rules refactor with my vacation in between19:04
*** xyang_ has joined #openstack-manila19:04
gansobswartz: so there are lots of things that I consider more important to finish and wrap up before updating the spec, as the code must be ready Jan 12th19:05
*** martyt has quit IRC19:06
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] debug 3.0  https://review.openstack.org/41036619:12
*** chlong has joined #openstack-manila19:13
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] debug 3.0  https://review.openstack.org/41036619:15
*** martyt has joined #openstack-manila19:17
*** dsariel has quit IRC19:20
*** ociuhandu has joined #openstack-manila19:22
*** xyang_ has quit IRC19:28
*** gouthamr has quit IRC19:29
*** rraja has quit IRC19:35
*** harlowja has quit IRC19:35
*** catintheroof has joined #openstack-manila19:37
*** catinthe_ has quit IRC19:40
*** martyt has quit IRC19:45
*** xyang_ has joined #openstack-manila19:48
*** gouthamr has joined #openstack-manila19:50
*** a-pugachev has joined #openstack-manila20:01
*** dsariel has joined #openstack-manila20:03
*** chlong has quit IRC20:15
*** breitz has quit IRC20:20
*** a-pugachev has quit IRC20:27
*** chlong has joined #openstack-manila20:27
*** dsariel has quit IRC20:28
*** cknight1 has joined #openstack-manila20:36
*** lpetrut has quit IRC20:36
cknight1ganso: ping20:36
*** lpetrut has joined #openstack-manila20:37
*** dsariel has joined #openstack-manila20:41
*** porrua has quit IRC20:47
*** catintheroof has quit IRC20:47
*** martyt has joined #openstack-manila20:51
openstackgerritClinton Knight proposed openstack/manila: Add create_share_from_snapshot_support extra spec  https://review.openstack.org/35668220:52
*** dsariel has quit IRC20:55
*** martyt has quit IRC20:56
*** breitz has joined #openstack-manila20:59
openstackgerritVitaliy Levitski proposed openstack/manila: Add MapR-FS native driver  https://review.openstack.org/35003521:01
*** alyson_ has quit IRC21:06
*** gouthamr has quit IRC21:09
*** lpetrut has quit IRC21:11
*** xyang_ has quit IRC21:12
*** mtanino has quit IRC21:13
*** breitz has quit IRC21:21
*** breitz_ has joined #openstack-manila21:21
*** sage__ has joined #openstack-manila21:35
*** nherciu has quit IRC21:36
openstackgerritMark Sturdevant proposed openstack/manila: GPFS KNFS: Do not reuse ssh prefix in loop  https://review.openstack.org/41323921:37
*** chlong has quit IRC21:46
*** akapil has joined #openstack-manila21:49
*** akapil has quit IRC21:54
*** cknight1 has quit IRC21:54
*** xyang_ has joined #openstack-manila22:08
*** xyang_ has quit IRC22:09
*** gouthamr has joined #openstack-manila22:13
*** xyang_ has joined #openstack-manila22:16
*** eharney has quit IRC22:21
*** mtanino has joined #openstack-manila22:21
*** xyang_ has quit IRC22:29
*** mtanino has quit IRC22:29
*** chlong has joined #openstack-manila22:57
*** dustins_ has quit IRC22:59
*** tommylikehu_ has quit IRC23:07
*** ociuhandu has quit IRC23:24
*** ociuhandu has joined #openstack-manila23:28
*** mtanino has joined #openstack-manila23:34
openstackgerritGoutham Pacha Ravi proposed openstack/manila: Add create_share_from_snapshot_support extra spec  https://review.openstack.org/35668223:48
*** ganso has quit IRC23:53
« Monday, 2016-12-19 Index Wednesday, 2016-12-21 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!