| hemna | no problemo :) | 00:00 |
|---|---|---|
| *** jerrygb has quit IRC | 00:01 | |
| *** zhangjn has joined #openstack-cinder | 00:05 | |
| *** zhangjn has quit IRC | 00:06 | |
| *** p0rtal_ has joined #openstack-cinder | 00:06 | |
| *** zhangjn has joined #openstack-cinder | 00:06 | |
| *** p0rtal has quit IRC | 00:09 | |
| *** diogogmt has quit IRC | 00:09 | |
| *** lcurtis has quit IRC | 00:10 | |
| *** Apoorva has joined #openstack-cinder | 00:12 | |
| *** salv-orl_ has joined #openstack-cinder | 00:18 | |
| *** akshai has quit IRC | 00:18 | |
| *** bardia has quit IRC | 00:18 | |
| openstackgerrit | Merged openstack/python-cinderclient: Pass proxy environment variable to tox https://review.openstack.org/252198 | 00:19 |
| *** bardia has joined #openstack-cinder | 00:19 | |
| openstackgerrit | Merged openstack/python-cinderclient: Fix comma location in comment https://review.openstack.org/252108 | 00:19 |
| openstackgerrit | Anthony Lee proposed openstack/cinder: LeftHand: Implement un/manage snapshot support https://review.openstack.org/255015 | 00:19 |
| openstackgerrit | Merged openstack/python-cinderclient: Add optional argument to list subcommand https://review.openstack.org/244962 | 00:19 |
| *** pck_ has joined #openstack-cinder | 00:20 | |
| *** hemna has quit IRC | 00:20 | |
| *** salv-orlando has quit IRC | 00:20 | |
| *** markstur has quit IRC | 00:20 | |
| *** hemna has joined #openstack-cinder | 00:20 | |
| *** markstur has joined #openstack-cinder | 00:20 | |
| *** pck has quit IRC | 00:22 | |
| *** zul has quit IRC | 00:22 | |
| *** jungleboyj has joined #openstack-cinder | 00:22 | |
| *** mordred has quit IRC | 00:23 | |
| *** zul has joined #openstack-cinder | 00:23 | |
| *** mordred has joined #openstack-cinder | 00:23 | |
| *** EinstCra_ has quit IRC | 00:24 | |
| *** zhangjn has quit IRC | 00:24 | |
| *** mc_nair_ has quit IRC | 00:26 | |
| *** chlong has joined #openstack-cinder | 00:28 | |
| *** takedakn has quit IRC | 00:30 | |
| *** mkoderer has quit IRC | 00:30 | |
| *** belmoreira has quit IRC | 00:30 | |
| *** alonma has joined #openstack-cinder | 00:33 | |
| mtanino | xyang: Hi, Do you think this patch is right fix? https://review.openstack.org/#/c/251713/ | 00:35 |
| mtanino | xyang: I think the fix conflicts your snapshot backup proposal. | 00:36 |
| *** mkoderer has joined #openstack-cinder | 00:37 | |
| *** alonma has quit IRC | 00:38 | |
| hemna | mtanino, xyang, what does the comment "when changing data | 00:39 |
| hemna | of a volume" mean ? | 00:39 |
| *** alonma has joined #openstack-cinder | 00:39 | |
| mtanino | hemna: xyang I guess that means "restore". | 00:42 |
| hemna | I guess. the commit message should be clarified | 00:43 |
| *** alonma has quit IRC | 00:44 | |
| mtanino | As snapshot is dependent on a volume, we should check whther the original volume has snapshots or not when we restore a backup. | 00:44 |
| hemna | yah that's better :) | 00:44 |
| mtanino | I imagine. | 00:44 |
| hemna | ok I'm out for the night... | 00:44 |
| *** hemna is now known as hemnafk | 00:44 | |
| mtanino | bye :) | 00:44 |
| tbarron | mtanino, lixiaoy1, xyang : so the idea here is that b/c snapshots are just deltas with the original volume, overwriting the original volume by an over-write restore would break that relationship, right? | 00:46 |
| tbarron | mtanino: you are however thinking that forbidding a restore in this circumstance conflicts with xyang's proposal to backup snapshots themselves? | 00:48 |
| *** openstackstatus has quit IRC | 00:50 | |
| *** openstack has joined #openstack-cinder | 00:52 | |
| mtanino | tbarron: But adding the check from that patch, user need to delete snapshot before restoring. | 00:52 |
| tbarron | mtanino, xyang, lixiaoy1: I havent' really thought through the implications of incremental backup and snapshot backup. | 00:52 |
| mtanino | tbarron: Ah, ok. | 00:52 |
| tbarron | mtanino, lixiaoy1, xyang: xyang has her hand in both of these areas, so I will wait to see what she says. Again, mtanino, good that you brought up the issue in any case. | 00:54 |
| *** zhangjn has joined #openstack-cinder | 00:55 | |
| mtanino | tbarron: yes, I agree. Waiting xyang's idea. | 00:55 |
| *** ianbrown_ has quit IRC | 01:03 | |
| *** EinstCrazy has joined #openstack-cinder | 01:03 | |
| *** zhangjn has quit IRC | 01:04 | |
| tbarron | somewhat related, I've recently had the (perhaps crazy) idea that we should not insist that an original volume cannot be deleted or overwritten if it has snapshots. | 01:05 |
| tbarron | it is true that some drivers implement snapshots as deltas to an original. | 01:06 |
| *** zhangjn has joined #openstack-cinder | 01:06 | |
| tbarron | but others do read-only clones. | 01:07 |
| tbarron | or some might do a dumb copy. | 01:07 |
| tbarron | so core code could allow the delete (from DB, etc.) and those drivers that do a delta from an original could keep the backing entity for the origianl | 01:07 |
| tbarron | around on their backend if they need to. | 01:08 |
| tbarron | why impose their way of implementation on the core code? | 01:08 |
| *** shihanzhang has joined #openstack-cinder | 01:15 | |
| lixiaoy1 | tbarron: mtanino: I agree that we should not insist that an original volume cannot be deleted or overwritten. it depends on whether it is full snapshot or incremental snapshot in backend | 01:17 |
| mtanino | tbarron: lixiaoy1 As for the delete case, I'm not sure the background why we can't delete it. | 01:18 |
| lixiaoy1 | tbarron:mtanino: if it is incremental snapshot, we can't delete it | 01:18 |
| mtanino | lixiaoy1: tbarron yes, but driver knows it. | 01:19 |
| *** shz has quit IRC | 01:19 | |
| lixiaoy1 | tbarron:mtanino: Iam not sure if it allows to delete volumes when it has snapshots, whether any problems happen | 01:20 |
| *** diogogmt has joined #openstack-cinder | 01:21 | |
| mtanino | lixiaoy1: tbarron So all drivers already depends on the current restriction, we can't change the policy. | 01:21 |
| mtanino | for delete case. | 01:22 |
| *** krtaylor has quit IRC | 01:22 | |
| lixiaoy1 | tbarron:mtanino: no | 01:22 |
| lixiaoy1 | tbarron:mtanino: I am not sure why at first we have such restriction | 01:22 |
| mtanino | lixiaoy1: tbarron ah, yes, me too | 01:23 |
| lixiaoy1 | tbarron:mtanino: meanwhile, I don't know how snapshots are used in cinder, we only have snapshot create, delete | 01:23 |
| mtanino | lixiaoy1: tbarron We can create a volume from snapshot | 01:23 |
| *** zhenguo has joined #openstack-cinder | 01:24 | |
| lixiaoy1 | tbarron:mtanino: oh, yeah | 01:24 |
| mtanino | lixiaoy1: tbarron It's main use case of snapshot , and we can backup snapshot in the future. | 01:25 |
| lixiaoy1 | tbarron:mtanino: let me check whether cinder copies all the metadata of a volume to its snapshot when creating snapshot, if yes, seems we can delete volumes if full snapshot | 01:25 |
| lixiaoy1 | tbarron:mtanino: very appreciate your help on this patch | 01:26 |
| mtanino | lixiaoy1: sure | 01:26 |
| *** garthb has quit IRC | 01:32 | |
| lixiaoy1 | tbarron:mtanino: I think not copy all the volume metadata during creating snapshot. | 01:33 |
| lixiaoy1 | tbarron:mtanino: this is in create_snapshot_in_db in volume.manager | 01:34 |
| lixiaoy1 | mtanino:as a result if deleting the volume, such metadata info lost | 01:35 |
| lixiaoy1 | mtanino: meanwhile same issue for extending volumes I think | 01:36 |
| *** krtaylor has joined #openstack-cinder | 01:36 | |
| *** zhangjn_ has joined #openstack-cinder | 01:36 | |
| mtanino | lixiaoy1: Why is it issue? | 01:38 |
| lixiaoy1 | mtanino: which one? metadata or extend? | 01:38 |
| mtanino | lixiaoy1: about metadata | 01:39 |
| lixiaoy1 | mtanino: when creating a volume from snapshot, we need to recover all the metadata | 01:39 |
| *** zhangjn has quit IRC | 01:40 | |
| *** Lee1092 has joined #openstack-cinder | 01:40 | |
| lixiaoy1 | mtanino: if a snapshot is full snapshot, and we delete the volume, we can't recover all the metadata except we save metadata in snapshot | 01:40 |
| *** alonma has joined #openstack-cinder | 01:41 | |
| mtanino | lixiaoy1: hmm, actually I don't have much knowledge around metadata, but it might be a design or problem. | 01:45 |
| lixiaoy1 | mtanino: I am not sure :) | 01:45 |
| *** alonma has quit IRC | 01:45 | |
| mtanino | lixiaoy1: Using volume clone(create --source-volid), it might copy full metadata | 01:46 |
| lixiaoy1 | mtanino: yes | 01:46 |
| *** esp has joined #openstack-cinder | 01:46 | |
| *** jamielennox is now known as jamielennox|away | 01:47 | |
| lixiaoy1 | mtanino: when creating a snapshot from a volume, it doesn't copy full metadata. as a result the snapshot depends on volumes. this is current mechanism | 01:47 |
| *** xyang has quit IRC | 01:48 | |
| lixiaoy1 | anyone can talk about snapshot usages in some drivers? | 01:48 |
| tbarron | lixiaoy1, mtanino : so you are saying the dependency of snapshot on original volume isn't due to backend implementations where it is a delta that depends on the original, but | 01:49 |
| tbarron | rather that the core code has chosen not to copy full metadata, right? | 01:50 |
| lixiaoy1 | mtanino: metadata is one aspect, another point is whether it is incremental or not in backend. | 01:52 |
| openstackgerrit | Dave McCowan proposed openstack/cinder: Check context before returning cached value https://review.openstack.org/254357 | 01:53 |
| lixiaoy1 | mtanino: if we change to use driver interface to decide whether depends or not, metadata needs to change | 01:53 |
| openstackgerrit | Merged openstack/cinder: XtremIO: fix missing multiattach flag https://review.openstack.org/251768 | 01:53 |
| *** leeantho has quit IRC | 01:54 | |
| tbarron | lixiaoy1, mtanino : yeah, but if xyang's proposal to backup from snapshot goes through (and I like it) then the base for incrmental could itself be a snapshot, righ? | 01:54 |
| tbarron | right? | 01:54 |
| lixiaoy1 | tbarron: yes | 01:54 |
| tbarron | incremental on base dependencies are one dimension, snap on original is another. | 01:55 |
| tbarron | I think the first is legitimate but am questioning the second. | 01:55 |
| lixiaoy1 | why? | 01:56 |
| tbarron | maybe snapshots should be from a core perspective be considered read-only clones/copies. | 01:56 |
| tbarron | if backends optimize them by doing deltas to an original, then that is their own concern, not core code's. | 01:56 |
| tbarron | lixiaoy1: because in general the core code doesn't have to keep an original in existence just becaue it has a snapshot. | 01:57 |
| tbarron | lixiaoy1: keeping it in existence because it has incremental backups is different. | 01:57 |
| tbarron | lixiaoy1: of course I said the last one wrong, keeping the base backup in existence because it has incremental is differnt. | 01:58 |
| tbarron | lixiaoy1: I might be confused, it's late here :-) | 01:58 |
| lixiaoy1 | tbarron: I am also confused about snapshot although it is morning here. I begin to notice my patch may be wrong | 02:00 |
| lixiaoy1 | tbarron: as during incremental snapshot, if the volume data are changed, the original data(dela) will be copied to snapshot if incremental | 02:01 |
| tbarron | lixiaoy1: I think your patch is consistent with current practice. mtanino has however raised good issues that lead me to wonder about the current practice. | 02:01 |
| tbarron | lixiaoy1: some backends might be safe if the original changes when they have done an incremental snapshot; some might not be and need to keep a shadow copy of the original around. | 02:02 |
| *** chlong has quit IRC | 02:03 | |
| *** xyang has joined #openstack-cinder | 02:03 | |
| tbarron | lixiaoy1: my thinking at the moment is that is the backend's concern, not the core code's. | 02:03 |
| lixiaoy1 | tbarron: but restoring the volume seems like to ios to the volume. we allow to write to the volume | 02:03 |
| lixiaoy1 | tbarron: so do you agree to use driver interface to check depencies like mtanino proposed? | 02:04 |
| tbarron | lixiaoy1: you are right, so I guess everyone has to handle that in some way or another on their backend. | 02:04 |
| tbarron | lixiaoy1: well, it is late :-), but I am more inclined to say backends should take care of any dependencies they have themselves, rather than | 02:05 |
| tbarron | the core code having to care. | 02:05 |
| mtanino | lixiaoy1: tbarron As for the backup case, I think above comment might be right solution. | 02:06 |
| lixiaoy1 | tbarron: thanks, then goodbye? | 02:06 |
| tbarron | Core code should be able to treat the backend as if it was doing a dumb point-in-time copy when doing a snapshot. | 02:06 |
| mtanino | nn, I might be confused. please ignore it. | 02:07 |
| tbarron | lixiaoy1: oh, I'll hang out a bit, but I'll shut up and listen :-) | 02:08 |
| lixiaoy1 | mtanino: tbarron: I would like to send an email to discuss the problem, ok? | 02:10 |
| tbarron | lixiaoy1, mtanino: violating my promise :-), OK by me. | 02:11 |
| lixiaoy1 | to check any other potential problems | 02:11 |
| mtanino | lixiaoy1: tbarron OK or good topic to discuss on the openstack-dev ML. | 02:12 |
| *** vgridnev has quit IRC | 02:12 | |
| tbarron | mtanino, lixiaoy1 good topic for ML I think. | 02:13 |
| lixiaoy1 | thanks, I will send an email | 02:13 |
| mtanino | so, good night! | 02:13 |
| lixiaoy1 | good night, you two | 02:13 |
| tbarron | lixiaoy1, mtanino: good "night" | 02:14 |
| *** Apoorva has quit IRC | 02:15 | |
| *** Apoorva has joined #openstack-cinder | 02:16 | |
| *** mtanino has quit IRC | 02:17 | |
| openstackgerrit | Wilson Liu proposed openstack/cinder: Rename Huawei drivers https://review.openstack.org/238041 | 02:20 |
| *** dave-mccowan has quit IRC | 02:20 | |
| openstackgerrit | Merged openstack/cinder: 3PAR: Implement v2 replication (managed) https://review.openstack.org/231201 | 02:25 |
| *** p0rtal has joined #openstack-cinder | 02:33 | |
| *** p0rtal_ has quit IRC | 02:37 | |
| openstackgerrit | Merged openstack/cinder: Volume driver for Coho Data storage solutions https://review.openstack.org/246690 | 02:40 |
| *** alonma has joined #openstack-cinder | 02:41 | |
| openstackgerrit | Merged openstack/cinder: Dynamically Pull Out Option Sections https://review.openstack.org/253277 | 02:41 |
| *** alonma has quit IRC | 02:45 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/253753 | 02:48 |
| *** jamielennox|away is now known as jamielennox | 02:50 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/python-cinderclient: Updated from global requirements https://review.openstack.org/250967 | 02:52 |
| openstackgerrit | wanghao proposed openstack/cinder-specs: Get the replica volume ref for DR https://review.openstack.org/182150 | 02:52 |
| *** diogogmt has quit IRC | 02:53 | |
| *** jerrygb has joined #openstack-cinder | 02:57 | |
| *** gouthamr has joined #openstack-cinder | 02:58 | |
| *** jerrygb has quit IRC | 03:02 | |
| *** jdurgin1 has quit IRC | 03:07 | |
| *** gouthamr_ has joined #openstack-cinder | 03:08 | |
| *** gouthamr has quit IRC | 03:11 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Storwize: Implement v2 replication (mirror) https://review.openstack.org/249042 | 03:13 |
| *** Apoorva has quit IRC | 03:14 | |
| *** chlong has joined #openstack-cinder | 03:16 | |
| *** jerrygb has joined #openstack-cinder | 03:18 | |
| *** links has joined #openstack-cinder | 03:24 | |
| *** mtanino has joined #openstack-cinder | 03:31 | |
| *** bardia has quit IRC | 03:31 | |
| itzdilip | how should be the multi pool is handled in driver? preconfigured with set of pools or creating on fly? | 03:34 |
| lixiaoy1 | itzdilip: as I know, some in conf file, some in extra_specs | 03:40 |
| lixiaoy1 | itzdilip: seems preconfigured | 03:41 |
| *** alonma has joined #openstack-cinder | 03:41 | |
| *** alonma has quit IRC | 03:46 | |
| *** jerrygb has quit IRC | 03:48 | |
| *** jerrygb has joined #openstack-cinder | 03:48 | |
| *** chlong has quit IRC | 03:49 | |
| *** alonma has joined #openstack-cinder | 03:51 | |
| *** alejandrito has quit IRC | 03:52 | |
| *** jerrygb has quit IRC | 03:52 | |
| *** chlong has joined #openstack-cinder | 03:53 | |
| *** [1]Thelo has joined #openstack-cinder | 03:55 | |
| *** alonma has quit IRC | 03:56 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Storwize: Implement v2 replication (mirror) https://review.openstack.org/249042 | 03:57 |
| *** alonma has joined #openstack-cinder | 03:57 | |
| *** [2]Thelo has quit IRC | 03:57 | |
| *** zhangjn_ has quit IRC | 03:58 | |
| *** xyang has quit IRC | 03:58 | |
| *** changbl has joined #openstack-cinder | 04:00 | |
| *** thambirajan has joined #openstack-cinder | 04:00 | |
| *** alonma has quit IRC | 04:02 | |
| *** links has quit IRC | 04:03 | |
| *** jerrygb has joined #openstack-cinder | 04:03 | |
| *** mtanino has quit IRC | 04:03 | |
| *** bardia has joined #openstack-cinder | 04:03 | |
| *** alonma has joined #openstack-cinder | 04:04 | |
| *** mattymo has quit IRC | 04:05 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Storwize: Implement v2 replication (mirror) https://review.openstack.org/249042 | 04:07 |
| *** alonma has quit IRC | 04:08 | |
| *** mattymo has joined #openstack-cinder | 04:11 | |
| *** gouthamr_ has quit IRC | 04:12 | |
| *** dims has joined #openstack-cinder | 04:13 | |
| *** dims has quit IRC | 04:15 | |
| *** dims has joined #openstack-cinder | 04:18 | |
| *** bardia has quit IRC | 04:19 | |
| *** changbl has quit IRC | 04:21 | |
| openstackgerrit | xing-yang proposed openstack/python-cinderclient: CLI for backup snapshots https://review.openstack.org/245632 | 04:22 |
| *** alonma has joined #openstack-cinder | 04:23 | |
| *** alonma has quit IRC | 04:27 | |
| *** klkumar has joined #openstack-cinder | 04:28 | |
| *** akshai has joined #openstack-cinder | 04:28 | |
| *** zhangjn has joined #openstack-cinder | 04:29 | |
| *** dims has quit IRC | 04:29 | |
| *** alonma has joined #openstack-cinder | 04:32 | |
| *** changbl has joined #openstack-cinder | 04:35 | |
| *** gouthamr has joined #openstack-cinder | 04:36 | |
| *** alonma has quit IRC | 04:37 | |
| *** links has joined #openstack-cinder | 04:38 | |
| *** gouthamr has quit IRC | 04:48 | |
| *** jerrygb has quit IRC | 04:49 | |
| *** jerrygb has joined #openstack-cinder | 04:51 | |
| *** jerrygb has quit IRC | 04:51 | |
| *** jerrygb has joined #openstack-cinder | 04:51 | |
| *** zhangjn has quit IRC | 04:52 | |
| *** boris-42_ has quit IRC | 05:03 | |
| *** p0rtal has quit IRC | 05:05 | |
| *** bardia has joined #openstack-cinder | 05:06 | |
| *** shausy has joined #openstack-cinder | 05:06 | |
| *** smoriya_ has joined #openstack-cinder | 05:09 | |
| *** smoriya has quit IRC | 05:10 | |
| *** klkumar has quit IRC | 05:11 | |
| *** sheel has joined #openstack-cinder | 05:14 | |
| *** klkumar has joined #openstack-cinder | 05:15 | |
| *** zhangjn has joined #openstack-cinder | 05:15 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Storwize: Implement v2 replication (mirror) https://review.openstack.org/249042 | 05:16 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Storwize: Implement v2 replication (split IO) https://review.openstack.org/237387 | 05:16 |
| *** harlowja_at_home has joined #openstack-cinder | 05:19 | |
| *** dims has joined #openstack-cinder | 05:20 | |
| *** dilip has joined #openstack-cinder | 05:24 | |
| *** dilip has quit IRC | 05:24 | |
| *** dilip has joined #openstack-cinder | 05:25 | |
| *** itzdilip has quit IRC | 05:26 | |
| *** klkumar has quit IRC | 05:26 | |
| *** dilip has quit IRC | 05:26 | |
| *** dilip has joined #openstack-cinder | 05:26 | |
| *** dilip has quit IRC | 05:26 | |
| *** dilip has joined #openstack-cinder | 05:27 | |
| *** dilip has quit IRC | 05:27 | |
| *** dilip has joined #openstack-cinder | 05:27 | |
| *** zhangjn has quit IRC | 05:28 | |
| *** zhangjn has joined #openstack-cinder | 05:28 | |
| *** dilip has quit IRC | 05:30 | |
| *** dilip has joined #openstack-cinder | 05:31 | |
| *** klkumar has joined #openstack-cinder | 05:32 | |
| *** martyturner has joined #openstack-cinder | 05:36 | |
| *** martyturner has quit IRC | 05:41 | |
| *** vgridnev has joined #openstack-cinder | 05:42 | |
| *** shausy has quit IRC | 05:44 | |
| *** shausy has joined #openstack-cinder | 05:44 | |
| *** alonma has joined #openstack-cinder | 05:46 | |
| *** alonma has quit IRC | 05:50 | |
| openstackgerrit | Merged openstack/cinder: Update migrate_volume API to use versionedobjects https://review.openstack.org/228916 | 05:52 |
| *** alonma has joined #openstack-cinder | 05:52 | |
| *** alonma has quit IRC | 05:57 | |
| *** salv-orl_ has quit IRC | 06:00 | |
| *** salv-orlando has joined #openstack-cinder | 06:00 | |
| *** alonma has joined #openstack-cinder | 06:02 | |
| *** alonma has quit IRC | 06:06 | |
| *** jerrygb has quit IRC | 06:17 | |
| *** jerrygb has joined #openstack-cinder | 06:17 | |
| *** salv-orl_ has joined #openstack-cinder | 06:17 | |
| *** alonma has joined #openstack-cinder | 06:17 | |
| *** salv-orlando has quit IRC | 06:20 | |
| *** dilip has quit IRC | 06:20 | |
| *** vinayp has joined #openstack-cinder | 06:21 | |
| *** dilip has joined #openstack-cinder | 06:21 | |
| *** dilip has quit IRC | 06:21 | |
| *** jerrygb has quit IRC | 06:21 | |
| *** itzdilip has joined #openstack-cinder | 06:22 | |
| *** alonma has quit IRC | 06:22 | |
| *** alonma has joined #openstack-cinder | 06:23 | |
| *** logan- has quit IRC | 06:24 | |
| *** sgotliv_ has joined #openstack-cinder | 06:26 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Imported Translations from Zanata https://review.openstack.org/255091 | 06:26 |
| *** ebalduf has quit IRC | 06:28 | |
| *** alonma has quit IRC | 06:28 | |
| *** alonma has joined #openstack-cinder | 06:30 | |
| *** jaypipes has joined #openstack-cinder | 06:31 | |
| *** alonma has quit IRC | 06:34 | |
| *** ChubYann has quit IRC | 06:37 | |
| *** lpetrut has joined #openstack-cinder | 06:37 | |
| *** chlong has quit IRC | 06:40 | |
| openstackgerrit | Rui Chen proposed openstack/python-cinderclient: Fix Resource.__eq__ mismatch semantics of object equal https://review.openstack.org/251021 | 06:44 |
| *** lpetrut has quit IRC | 06:48 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Add cluster Pairing setup to SolidFire driver https://review.openstack.org/238235 | 06:49 |
| *** shausy has quit IRC | 06:50 | |
| *** shausy has joined #openstack-cinder | 06:50 | |
| *** nkrinner has joined #openstack-cinder | 06:56 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Add replicate by volume-type to SolidFire driver https://review.openstack.org/238276 | 06:59 |
| openstackgerrit | John Griffith proposed openstack/cinder: Add cluster Pairing setup to SolidFire driver https://review.openstack.org/238235 | 06:59 |
| *** jerrygb has joined #openstack-cinder | 07:00 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/253753 | 07:04 |
| openstackgerrit | John Griffith proposed openstack/cinder: Replace use of mox with mock in test_solidfire https://review.openstack.org/255103 | 07:05 |
| *** jerrygb has quit IRC | 07:06 | |
| *** harlowja_at_home has quit IRC | 07:06 | |
| *** bardia has quit IRC | 07:08 | |
| *** rcernin has joined #openstack-cinder | 07:20 | |
| *** boris-42_ has joined #openstack-cinder | 07:31 | |
| *** sgotliv_ has quit IRC | 07:34 | |
| *** zhangjn has quit IRC | 07:36 | |
| *** itzdilip has quit IRC | 07:37 | |
| *** alexschm has joined #openstack-cinder | 07:38 | |
| *** itzdilip has joined #openstack-cinder | 07:38 | |
| *** klkumar has quit IRC | 07:41 | |
| *** alonma has joined #openstack-cinder | 07:42 | |
| *** itzdilip has quit IRC | 07:42 | |
| *** klkumar has joined #openstack-cinder | 07:44 | |
| *** itzdilip has joined #openstack-cinder | 07:45 | |
| *** itzdilip has quit IRC | 07:45 | |
| *** itzdilip has joined #openstack-cinder | 07:46 | |
| *** nkrinner has quit IRC | 07:46 | |
| *** dilip has joined #openstack-cinder | 07:46 | |
| *** nkrinner has joined #openstack-cinder | 07:46 | |
| *** alonma has quit IRC | 07:46 | |
| *** itzdilip has quit IRC | 07:48 | |
| *** rhagarty_ has joined #openstack-cinder | 07:48 | |
| *** rhagarty has quit IRC | 07:49 | |
| *** e0ne has joined #openstack-cinder | 07:51 | |
| *** alonma has joined #openstack-cinder | 07:52 | |
| *** yhayashi has quit IRC | 07:52 | |
| *** dilip has quit IRC | 07:53 | |
| *** dilip has joined #openstack-cinder | 07:53 | |
| *** nkrinner has quit IRC | 07:54 | |
| *** nkrinner has joined #openstack-cinder | 07:55 | |
| *** alonma has quit IRC | 07:57 | |
| openstackgerrit | wanghao proposed openstack/cinder: Get the replica volume ref for DR https://review.openstack.org/252250 | 07:59 |
| *** alonma has joined #openstack-cinder | 07:59 | |
| *** dilip has quit IRC | 07:59 | |
| openstackgerrit | wanghao proposed openstack/cinder: Get the replica volume ref for DR https://review.openstack.org/252250 | 07:59 |
| *** zhangjn has joined #openstack-cinder | 08:00 | |
| *** alonma has quit IRC | 08:03 | |
| *** anshul has joined #openstack-cinder | 08:04 | |
| openstackgerrit | wanghao proposed openstack/cinder-specs: Get the replica volume ref for DR https://review.openstack.org/182150 | 08:04 |
| *** alonma has joined #openstack-cinder | 08:05 | |
| *** alonma has quit IRC | 08:05 | |
| *** egonzalez has joined #openstack-cinder | 08:06 | |
| *** alonma has joined #openstack-cinder | 08:06 | |
| *** isaacb has joined #openstack-cinder | 08:06 | |
| *** dilip has joined #openstack-cinder | 08:13 | |
| *** liverpooler has joined #openstack-cinder | 08:15 | |
| *** belmoreira has joined #openstack-cinder | 08:15 | |
| *** klkumar has quit IRC | 08:17 | |
| *** vgridnev has quit IRC | 08:23 | |
| *** markus_z has joined #openstack-cinder | 08:23 | |
| *** smoriya_ has quit IRC | 08:31 | |
| *** salv-orl_ has quit IRC | 08:58 | |
| *** jordanP has joined #openstack-cinder | 08:58 | |
| *** jerrygb has joined #openstack-cinder | 09:02 | |
| *** openstackstatus has quit IRC | 09:05 | |
| *** openstack has joined #openstack-cinder | 09:07 | |
| -wolfe.freenode.net- [freenode-info] channel flooding and no channel staff around to help? Please check with freenode support: http://freenode.net/faq.shtml#gettinghelp | 09:07 | |
| *** jerrygb has quit IRC | 09:08 | |
| *** alonma has joined #openstack-cinder | 09:09 | |
| *** ociuhandu has quit IRC | 09:14 | |
| *** e0ne has quit IRC | 09:15 | |
| *** dilip has quit IRC | 09:16 | |
| *** dilip has joined #openstack-cinder | 09:17 | |
| *** sgotliv_ has joined #openstack-cinder | 09:18 | |
| *** chenying has joined #openstack-cinder | 09:19 | |
| *** skraynev has quit IRC | 09:20 | |
| *** chenying_ has quit IRC | 09:21 | |
| *** ndipanov has joined #openstack-cinder | 09:27 | |
| *** shausy has quit IRC | 09:28 | |
| *** skraynev has joined #openstack-cinder | 09:28 | |
| *** shausy has joined #openstack-cinder | 09:29 | |
| *** alonma has quit IRC | 09:31 | |
| *** vgridnev has joined #openstack-cinder | 09:32 | |
| *** openstackgerrit has quit IRC | 09:32 | |
| *** openstackgerrit has joined #openstack-cinder | 09:33 | |
| *** alonma has joined #openstack-cinder | 09:33 | |
| *** shausy has quit IRC | 09:42 | |
| *** ildikov has quit IRC | 09:42 | |
| *** shausy has joined #openstack-cinder | 09:42 | |
| *** zhangjn has quit IRC | 09:45 | |
| *** skraynev has quit IRC | 09:48 | |
| *** skraynev has joined #openstack-cinder | 09:49 | |
| *** dilip has quit IRC | 09:50 | |
| *** dilip has joined #openstack-cinder | 09:50 | |
| *** dilip has quit IRC | 09:50 | |
| *** itzdilip has joined #openstack-cinder | 09:51 | |
| *** zhangjn has joined #openstack-cinder | 09:57 | |
| *** salv-orlando has joined #openstack-cinder | 09:59 | |
| *** openstackgerrit has quit IRC | 10:02 | |
| *** gcb has quit IRC | 10:02 | |
| *** zhenguo has quit IRC | 10:02 | |
| *** openstackgerrit has joined #openstack-cinder | 10:03 | |
| *** salv-orlando has quit IRC | 10:03 | |
| *** ildikov has joined #openstack-cinder | 10:09 | |
| *** e0ne has joined #openstack-cinder | 10:10 | |
| *** jamielennox is now known as jamielennox|away | 10:13 | |
| *** sileht has quit IRC | 10:16 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Add synchronization in Block Device driver https://review.openstack.org/252979 | 10:18 |
| *** sileht has joined #openstack-cinder | 10:19 | |
| *** sileht has quit IRC | 10:19 | |
| *** sileht has joined #openstack-cinder | 10:20 | |
| *** aix has joined #openstack-cinder | 10:20 | |
| *** rushiagr_away is now known as rushiagr | 10:21 | |
| *** dims has quit IRC | 10:25 | |
| *** alexschm has quit IRC | 10:26 | |
| *** alexschm has joined #openstack-cinder | 10:29 | |
| *** ianbrown has joined #openstack-cinder | 10:30 | |
| *** alonma has quit IRC | 10:31 | |
| *** alonma has joined #openstack-cinder | 10:31 | |
| *** alexschm_ has joined #openstack-cinder | 10:31 | |
| *** cbader has quit IRC | 10:32 | |
| *** cbader has joined #openstack-cinder | 10:33 | |
| *** alexschm has quit IRC | 10:34 | |
| *** alonma has quit IRC | 10:35 | |
| *** aix has quit IRC | 10:37 | |
| *** alonma has joined #openstack-cinder | 10:37 | |
| *** jistr has joined #openstack-cinder | 10:39 | |
| *** alonma has quit IRC | 10:41 | |
| *** eduardo___ has joined #openstack-cinder | 10:44 | |
| *** alonma has joined #openstack-cinder | 10:59 | |
| *** alonma has quit IRC | 11:04 | |
| *** jerrygb has joined #openstack-cinder | 11:04 | |
| *** kmartin has quit IRC | 11:05 | |
| *** kmartin has joined #openstack-cinder | 11:05 | |
| *** alonma has joined #openstack-cinder | 11:06 | |
| *** jerrygb has quit IRC | 11:09 | |
| *** salv-orlando has joined #openstack-cinder | 11:10 | |
| *** alonma has quit IRC | 11:10 | |
| *** alonma has joined #openstack-cinder | 11:10 | |
| *** lprice has joined #openstack-cinder | 11:13 | |
| *** lprice1 has quit IRC | 11:13 | |
| *** aix has joined #openstack-cinder | 11:14 | |
| *** dilip has joined #openstack-cinder | 11:14 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Small refactoring in test_admin_actions.py https://review.openstack.org/253541 | 11:15 |
| *** itzdilip has quit IRC | 11:15 | |
| *** adrianofr has joined #openstack-cinder | 11:16 | |
| *** dilip has quit IRC | 11:16 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Small refactoring in test_admin_actions.py https://review.openstack.org/253541 | 11:16 |
| *** dilip has joined #openstack-cinder | 11:17 | |
| *** zhangjn has quit IRC | 11:23 | |
| *** EinstCrazy has quit IRC | 11:24 | |
| *** dilip has quit IRC | 11:25 | |
| *** abhi has quit IRC | 11:26 | |
| *** links has quit IRC | 11:26 | |
| *** klkumar has joined #openstack-cinder | 11:28 | |
| *** angela-s has quit IRC | 11:29 | |
| *** dims has joined #openstack-cinder | 11:37 | |
| *** dilip has joined #openstack-cinder | 11:39 | |
| *** rushiagr is now known as rushiagr_away | 11:48 | |
| *** ndipanov has quit IRC | 11:48 | |
| *** ndipanov has joined #openstack-cinder | 11:49 | |
| *** kmartin has quit IRC | 11:49 | |
| *** rushiagr_away is now known as rushiagr | 11:49 | |
| *** zhangjn has joined #openstack-cinder | 11:49 | |
| *** EinstCrazy has joined #openstack-cinder | 11:51 | |
| *** dilip has quit IRC | 11:52 | |
| *** rushiagr is now known as rushiagr_away | 11:52 | |
| *** cdelatte has quit IRC | 12:05 | |
| *** takedakn has joined #openstack-cinder | 12:11 | |
| *** Lee1092 has quit IRC | 12:11 | |
| *** links has joined #openstack-cinder | 12:16 | |
| *** salv-orl_ has joined #openstack-cinder | 12:17 | |
| *** salv-orlando has quit IRC | 12:20 | |
| *** smoriya has joined #openstack-cinder | 12:24 | |
| *** houming has joined #openstack-cinder | 12:31 | |
| *** rushiagr_away is now known as rushiagr | 12:32 | |
| *** gcb has joined #openstack-cinder | 12:33 | |
| *** alejandrito has joined #openstack-cinder | 12:40 | |
| *** chlong has joined #openstack-cinder | 12:42 | |
| *** boris-42_ has quit IRC | 12:43 | |
| *** dilip has joined #openstack-cinder | 12:43 | |
| *** salv-orl_ has quit IRC | 12:44 | |
| openstackgerrit | Kazumasa Nomura proposed openstack/cinder: Refactor Hitachi Block Storage Driver https://review.openstack.org/255239 | 12:44 |
| *** dilip has quit IRC | 12:46 | |
| *** cdelatte has joined #openstack-cinder | 12:48 | |
| *** takedakn1 has joined #openstack-cinder | 12:48 | |
| *** salv-orlando has joined #openstack-cinder | 12:49 | |
| *** takedakn has quit IRC | 12:49 | |
| *** takedakn1 is now known as takedakn | 12:49 | |
| *** jerrygb has joined #openstack-cinder | 12:53 | |
| *** nikeshm has joined #openstack-cinder | 12:55 | |
| nikeshm | hi | 12:55 |
| nikeshm | bp "https://blueprints.launchpad.net/cinder/+spec/cinder-python3" was registered and implemented in liberty, can we use same bp for mitaka for this patch https://review.openstack.org/#/c/254198/ | 12:57 |
| nikeshm | its already got one +2, so thought better to arise this | 12:57 |
| e0ne | nikeshm: it's partially implemented in Liberty, so it seems to be OK to re-use it instead of creating a new one | 12:58 |
| nikeshm | e0ne : hmm, ok , thanks | 12:59 |
| *** takedakn has quit IRC | 12:59 | |
| *** dilip has joined #openstack-cinder | 13:00 | |
| *** zhipeng has joined #openstack-cinder | 13:00 | |
| *** dilip has quit IRC | 13:01 | |
| *** julim has joined #openstack-cinder | 13:01 | |
| *** Lee1092 has joined #openstack-cinder | 13:06 | |
| *** ildikov has quit IRC | 13:07 | |
| *** asselin has joined #openstack-cinder | 13:19 | |
| *** ildikov has joined #openstack-cinder | 13:22 | |
| *** chenying has quit IRC | 13:29 | |
| *** chenying has joined #openstack-cinder | 13:30 | |
| *** edmondsw has joined #openstack-cinder | 13:36 | |
| *** martyturner has joined #openstack-cinder | 13:36 | |
| *** takedakn has joined #openstack-cinder | 13:37 | |
| *** gouthamr has joined #openstack-cinder | 13:40 | |
| *** akerr has joined #openstack-cinder | 13:40 | |
| *** zhonghua-lee has quit IRC | 13:40 | |
| *** zhonghua-lee has joined #openstack-cinder | 13:41 | |
| *** gouthamr_ has joined #openstack-cinder | 13:41 | |
| *** diogogmt has joined #openstack-cinder | 13:41 | |
| *** zhipeng has quit IRC | 13:42 | |
| *** dilip has joined #openstack-cinder | 13:42 | |
| *** eharney has joined #openstack-cinder | 13:44 | |
| *** gouthamr has quit IRC | 13:45 | |
| *** timcl has joined #openstack-cinder | 13:49 | |
| *** dilip has quit IRC | 13:52 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Attach/detach volumes without Nova https://review.openstack.org/224124 | 13:52 |
| *** itzdilip has joined #openstack-cinder | 13:52 | |
| *** Thelo has joined #openstack-cinder | 13:55 | |
| *** [1]Thelo has quit IRC | 13:57 | |
| *** logan- has joined #openstack-cinder | 14:01 | |
| openstackgerrit | Erik Johannes proposed openstack/cinder: Implement CHAP Authentication for E-Series Driver https://review.openstack.org/246990 | 14:02 |
| *** sgotliv_ has quit IRC | 14:03 | |
| *** martyturner has quit IRC | 14:03 | |
| *** gcb has quit IRC | 14:06 | |
| *** martyturner has joined #openstack-cinder | 14:08 | |
| *** jerrygb_ has joined #openstack-cinder | 14:08 | |
| *** houming has quit IRC | 14:09 | |
| *** haomaiwang has joined #openstack-cinder | 14:10 | |
| *** jerrygb has quit IRC | 14:11 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Add synchronization in Block Device driver https://review.openstack.org/252979 | 14:13 |
| *** houming has joined #openstack-cinder | 14:17 | |
| *** sgotliv_ has joined #openstack-cinder | 14:18 | |
| *** takedakn has quit IRC | 14:19 | |
| *** breitz has quit IRC | 14:19 | |
| *** breitz has joined #openstack-cinder | 14:19 | |
| *** merooney has joined #openstack-cinder | 14:21 | |
| *** diogogmt has quit IRC | 14:24 | |
| *** mriedem_away is now known as mriedem | 14:25 | |
| *** diogogmt has joined #openstack-cinder | 14:26 | |
| *** zhipeng has joined #openstack-cinder | 14:27 | |
| *** jgregor has joined #openstack-cinder | 14:27 | |
| *** shausy has quit IRC | 14:28 | |
| *** ildikov has quit IRC | 14:28 | |
| *** ildikov has joined #openstack-cinder | 14:30 | |
| *** diablo_rojo has joined #openstack-cinder | 14:31 | |
| *** mtanino has joined #openstack-cinder | 14:31 | |
| *** ildikov has quit IRC | 14:35 | |
| *** timwilliams has joined #openstack-cinder | 14:35 | |
| *** timwilliams has quit IRC | 14:37 | |
| *** bardia has joined #openstack-cinder | 14:38 | |
| *** diogogmt has quit IRC | 14:38 | |
| *** dave-mccowan has joined #openstack-cinder | 14:38 | |
| *** akshai_ has joined #openstack-cinder | 14:40 | |
| *** zhipeng has quit IRC | 14:42 | |
| *** akshai has quit IRC | 14:44 | |
| *** mc_nair has joined #openstack-cinder | 14:45 | |
| *** lpabon has joined #openstack-cinder | 14:45 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Add synchronization in Block Device driver https://review.openstack.org/252979 | 14:46 |
| openstackgerrit | Sean McGinnis proposed openstack/python-cinderclient: DO NOT MERGE: This is only a test https://review.openstack.org/255292 | 14:47 |
| anteaya | smcginnis: I see that | 14:47 |
| smcginnis | anteaya: OK, nevermind. Appears to be working now. Sorry for the distraction. :) | 14:47 |
| anteaya | no worries | 14:48 |
| anteaya | glad I had to do nothing to make you happy | 14:48 |
| anteaya | :) | 14:48 |
| anteaya | good start to the day | 14:48 |
| smcginnis | anteaya: It definitely wasn't the other day. I'll keep an eye on it. :) | 14:48 |
| smcginnis | anteaya: Yeah, good easy one for the morning. :D | 14:48 |
| anteaya | okay if you see it disappear to let me know | 14:48 |
| anteaya | now keep in mind freenode has had a bad time lately | 14:48 |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Implement snapshots-related features for Block Device Driver https://review.openstack.org/253111 | 14:48 |
| anteaya | if openstackgerrit isn't in channel it can't report anything | 14:48 |
| smcginnis | anteaya: Yeah, it's quite possible it was around the time of the ddos. | 14:49 |
| anteaya | that might explain the silence then | 14:49 |
| smcginnis | anteaya: It was odd though that we were getting notifications for cinder and os-brick, just not client. | 14:49 |
| smcginnis | Oh well. | 14:49 |
| anteaya | while we get grumpy when it happens fungi has to play server roulette to get our bots back in action | 14:50 |
| smcginnis | anteaya: Not fun, I'm sure! | 14:50 |
| *** ntpttr has joined #openstack-cinder | 14:51 | |
| anteaya | not so much | 14:52 |
| anteaya | I don't do anything | 14:52 |
| anteaya | but fungi does | 14:52 |
| *** merooney has quit IRC | 14:59 | |
| *** ndipanov has quit IRC | 15:02 | |
| *** eharney has quit IRC | 15:02 | |
| e0ne | jgriffith: hi. FYI, I've updated spec https://review.openstack.org/224124 to implement cinderclient extension instead of python-brickclient | 15:02 |
| *** ndipanov has joined #openstack-cinder | 15:02 | |
| *** dustins|gone is now known as dustins | 15:03 | |
| *** flaper87 has joined #openstack-cinder | 15:03 | |
| *** eharney has joined #openstack-cinder | 15:03 | |
| *** flaper87 has left #openstack-cinder | 15:05 | |
| *** merooney has joined #openstack-cinder | 15:05 | |
| *** xyang1 has joined #openstack-cinder | 15:06 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Implement snapshots-related features for Block Device Driver https://review.openstack.org/253111 | 15:08 |
| *** wanghao has joined #openstack-cinder | 15:09 | |
| jgriffith | e0ne: cool, I'll have a look. Tahnks! | 15:09 |
| *** cdelatte has quit IRC | 15:09 | |
| *** cdelatte has joined #openstack-cinder | 15:10 | |
| e0ne | jgriffith: great, thanks! there is also a draft implementation for it | 15:10 |
| *** mdenny has joined #openstack-cinder | 15:11 | |
| *** gouthamr_ has quit IRC | 15:11 | |
| wanghao | jgriffith:thingee,smcginnis | 15:11 |
| *** timcl1 has joined #openstack-cinder | 15:11 | |
| smcginnis | wanghao: Hey | 15:11 |
| wanghao | smcginnis: hi, I have updated this spec: get replica volume ref | 15:12 |
| smcginnis | wanghao: Already open in a tab. ;) | 15:12 |
| *** delattec has joined #openstack-cinder | 15:12 | |
| smcginnis | wanghao: I'm hoping to review it again today. | 15:12 |
| smcginnis | Just have a few things to get through first. | 15:12 |
| jgriffith | smcginnis: what... are you psychic! | 15:12 |
| wanghao | smcginnis: yeah, great, thanks | 15:12 |
| smcginnis | jgriffith: ;) | 15:13 |
| wanghao | smcginnis: jgriffith I was shocked a little :) | 15:14 |
| *** cdelatte has quit IRC | 15:15 | |
| *** timcl has quit IRC | 15:15 | |
| jgriffith | wanghao: :) | 15:15 |
| wanghao | thingee: hi~, are you here? | 15:16 |
| smcginnis | Still pretty early for him. He's US west coast. | 15:17 |
| smcginnis | He's pretty good about seeing pings while he's gone though. | 15:17 |
| wanghao | smcginnis:okay, got it. | 15:18 |
| *** dilip has joined #openstack-cinder | 15:19 | |
| *** itzdilip has quit IRC | 15:19 | |
| *** alonma has quit IRC | 15:19 | |
| wanghao | smcginnis:Just wanna to talked about his suggestion in this spec. I have a little confusion about it. | 15:20 |
| *** alonma has joined #openstack-cinder | 15:20 | |
| *** ildikov has joined #openstack-cinder | 15:20 | |
| *** ebalduf has joined #openstack-cinder | 15:22 | |
| *** noqa_v_qoovnie has quit IRC | 15:22 | |
| *** alonma has quit IRC | 15:24 | |
| *** diogogmt has joined #openstack-cinder | 15:25 | |
| *** alonma has joined #openstack-cinder | 15:26 | |
| *** alonma has quit IRC | 15:26 | |
| *** dilip has quit IRC | 15:27 | |
| *** alonma has joined #openstack-cinder | 15:27 | |
| jgriffith | wanghao: could I ask a few questions about this? | 15:28 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Move retype quota checks to API https://review.openstack.org/254980 | 15:29 |
| *** ociuhandu has joined #openstack-cinder | 15:30 | |
| wanghao | jgriffith: sure | 15:31 |
| openstackgerrit | Ryan McNair proposed openstack/cinder: Fix non-migration swap with error https://review.openstack.org/252089 | 15:31 |
| *** houming has quit IRC | 15:31 | |
| jgriffith | wanghao: I posted some comments in the review (well.. quesitons more than comments) | 15:32 |
| jgriffith | wanghao: can you hep me understand the use case for this a bit better? More specifically what problem you're trying to solve. | 15:33 |
| jgriffith | wanghao: I'm afraid I didn't quite figure it out in the spec, and the code ruins the assumption that I did have | 15:33 |
| *** anshul has quit IRC | 15:33 | |
| jgriffith | wanghao: because you're adding a volume_ref field, but I'm confused because it's a string and I don't know what it is supposed to "be" or what you intend to use it for | 15:34 |
| jgriffith | wanghao: my apologies for not "getting it" :) | 15:34 |
| *** dilip has joined #openstack-cinder | 15:34 | |
| *** gouthamr has joined #openstack-cinder | 15:35 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Attach/detach volumes without Nova https://review.openstack.org/224124 | 15:37 |
| *** timcl1 has quit IRC | 15:37 | |
| wanghao | jgriffith: okay, simplely to say, if there are two independent cloud sites, we create replication volume in site 1, and drivers build replica volume in site 2's storage. well, when disaster happen, site 1 is down, and then we want to find our data back and continue to use it in site 2. | 15:39 |
| *** belmoreira has quit IRC | 15:39 | |
| smcginnis | wanghao: So to validate my high level understanding so far: | 15:40 |
| openstackgerrit | John Griffith proposed openstack/cinder: Add replicate by volume-type to SolidFire driver https://review.openstack.org/238276 | 15:40 |
| wanghao | jgriffith: but how? So we want to manage the replica volume in site 2. | 15:40 |
| smcginnis | wanghao: This is for DR. If the primary goes down, we need to be able to feed the replication target information into the secondary side's Cinder service. | 15:40 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Move snapshot and volume quota checks to API https://review.openstack.org/249388 | 15:40 |
| smcginnis | wanghao: So something like a manage existing, but passing in details so the driver knows it's managing what was a replication target? | 15:40 |
| openstackgerrit | John Griffith proposed openstack/cinder: Add cluster Pairing setup to SolidFire driver https://review.openstack.org/238235 | 15:42 |
| jgriffith | wanghao: ok, so I kinda get what you're trying to accomplish, but I'm not following how this will help you accomplish it | 15:43 |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Attach/detach volumes without Nova https://review.openstack.org/224124 | 15:43 |
| jgriffith | wanghao: I'm also not sure how I feel about explicitly putting entries in the db to point to references in "another" cloud | 15:43 |
| jgriffith | wanghao: that can be a VERY slippery slope | 15:43 |
| *** thangp has joined #openstack-cinder | 15:43 | |
| smcginnis | Yeah, not sure we need to muck with the DB here. | 15:44 |
| jgriffith | wanghao: my point is I believe this can be solved other ways maybe, using mechanisms we already have in place | 15:44 |
| *** baumann has joined #openstack-cinder | 15:44 | |
| *** rs007 has joined #openstack-cinder | 15:44 | |
| jgriffith | wanghao: say like maybe admin-metadata... or even an explicit type | 15:44 |
| *** rs007_ has joined #openstack-cinder | 15:44 | |
| smcginnis | Probably just a get_repl_metadata() type call on the source side, then a manage_existing(repl_metadata) on the dest. | 15:44 |
| *** rs007 has quit IRC | 15:44 | |
| * smcginnis is brainstorming without having reread any recent details of what the spec actually calls for | 15:45 | |
| jgriffith | wanghao: because in the case of a cloud2cloud DR setup and admin can/should have some special things in place IMHO. One of those things could be a "replicated-volume" type | 15:45 |
| e0ne | jgriffith, smcginnis: thanks for the review! | 15:45 |
| jgriffith | wanghao: that in combination with some metadata perhaps would accomplish the same thing here no? | 15:45 |
| wanghao | jgriffith:smcginnis em... at first I want to put this into replication_driver_data | 15:46 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Move consistency group quota checks to API https://review.openstack.org/249441 | 15:48 |
| jgriffith | wanghao: so let's back up a second... can you provide me an example of "what" that column would be? | 15:50 |
| *** alexschm_ has quit IRC | 15:51 | |
| wanghao | jgriffith: In my mind, it will just be a string column, that drivers put the replica volume ref into it when they build the replication in source side. | 15:52 |
| *** zhangjn has quit IRC | 15:52 | |
| wanghao | jgriffith: Like replication_driver_data | 15:53 |
| jgriffith | wanghao: what do you mean by "replica_volume_ref" though? | 15:53 |
| jgriffith | wanghao: what would that actually "be"... a provider id? a cinder-uuid? | 15:54 |
| *** liverpooler has quit IRC | 15:55 | |
| wanghao | jgriffith: emm, more like provider id, you can use it to manage this volume in another openstack system. | 15:55 |
| *** sheel has quit IRC | 15:55 | |
| wanghao | jgriffith: Like if we want to manage a LVM volume, we need to know it's lv name. | 15:56 |
| jgriffith | wanghao: sure, so we could probably use existing rep data for that don't you think? | 15:56 |
| jgriffith | wanghao: or even tack something on to the provider_id field | 15:57 |
| jgriffith | wanghao: the name "volume_ref" is kinda misleading by the way | 15:57 |
| jgriffith | wanghao: so for example provider_id: rep_target_id: "foo" | 15:57 |
| *** haomaiwang has quit IRC | 15:57 | |
| jgriffith | or something like that | 15:57 |
| *** zhipeng has joined #openstack-cinder | 15:58 | |
| jgriffith | or back to using the replication_driver_data column | 15:58 |
| smcginnis | Meeting time | 15:58 |
| jgriffith | it seems to me like that would fit and kinda makes more sense | 15:58 |
| *** haomaiwang has joined #openstack-cinder | 15:58 | |
| *** jseiler has joined #openstack-cinder | 15:58 | |
| jgriffith | ok... meeting!!! Yay!!! | 15:58 |
| wanghao | jgriffith: yes, exactly | 15:58 |
| *** haomaiwang has quit IRC | 15:59 | |
| *** erlon has joined #openstack-cinder | 15:59 | |
| *** haomaiwa_ has joined #openstack-cinder | 15:59 | |
| *** haomaiwa_ has quit IRC | 16:00 | |
| *** rhedlind has joined #openstack-cinder | 16:00 | |
| *** haomaiwang has joined #openstack-cinder | 16:00 | |
| *** pots has joined #openstack-cinder | 16:00 | |
| wanghao | jgriffith: but I feel it's a little hard to handle it in replication_driver_data, we need to handle the string and get the 'provider_id:rep_target_id:"foo" ' from a long string. | 16:00 |
| *** haomaiwang has quit IRC | 16:01 | |
| *** haomaiwa_ has joined #openstack-cinder | 16:01 | |
| *** pck_ is now known as pck | 16:02 | |
| yuriy_n17 | join #openstack-meeting | 16:02 |
| *** yuriy_n17 has quit IRC | 16:03 | |
| *** zz_john5223 is now known as john5223 | 16:03 | |
| *** kmartin has joined #openstack-cinder | 16:07 | |
| *** nkrinner has quit IRC | 16:08 | |
| *** jerrygb_ has quit IRC | 16:18 | |
| *** kfarr has joined #openstack-cinder | 16:18 | |
| *** diogogmt has quit IRC | 16:22 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Small refactoring in test_admin_actions.py https://review.openstack.org/253541 | 16:22 |
| *** vgridnev has quit IRC | 16:23 | |
| *** alonma has quit IRC | 16:23 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Add synchronization in Block Device driver https://review.openstack.org/252979 | 16:26 |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Implement snapshots-related features for Block Device Driver https://review.openstack.org/253111 | 16:27 |
| *** harlowja_at_home has joined #openstack-cinder | 16:27 | |
| *** hemnafk is now known as hemna | 16:27 | |
| openstackgerrit | Mikhail Khodos proposed openstack/cinder: --amend https://review.openstack.org/255359 | 16:29 |
| *** martyturner has quit IRC | 16:30 | |
| *** alonma has joined #openstack-cinder | 16:30 | |
| *** crose has joined #openstack-cinder | 16:31 | |
| *** martyturner has joined #openstack-cinder | 16:33 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Attach/detach volumes without Nova https://review.openstack.org/224124 | 16:33 |
| *** crose has quit IRC | 16:33 | |
| *** alonma has quit IRC | 16:35 | |
| *** alonma has joined #openstack-cinder | 16:36 | |
| *** haomaiwa_ has quit IRC | 16:40 | |
| *** haomaiwang has joined #openstack-cinder | 16:40 | |
| *** alonma has quit IRC | 16:40 | |
| *** alonma has joined #openstack-cinder | 16:42 | |
| *** ctina_ has joined #openstack-cinder | 16:43 | |
| *** p0rtal has joined #openstack-cinder | 16:44 | |
| *** lpabon has quit IRC | 16:46 | |
| *** alonma has quit IRC | 16:47 | |
| *** markstur has quit IRC | 16:48 | |
| *** alonma has joined #openstack-cinder | 16:49 | |
| *** markstur has joined #openstack-cinder | 16:50 | |
| *** earlephilhower has joined #openstack-cinder | 16:50 | |
| *** klkumar has quit IRC | 16:51 | |
| *** alonma has quit IRC | 16:53 | |
| *** rushiagr is now known as rushiagr_away | 16:53 | |
| *** rushiagr_away is now known as rushiagr | 16:54 | |
| winston-d | e0ne: so you mean you didn't add that to cinderclient? | 16:59 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Move retype quota checks to API https://review.openstack.org/254980 | 17:00 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Move retype quota checks to API https://review.openstack.org/254980 | 17:00 |
| e0ne | winston-d: I've got a PoC implementation with cinderclient and get-conector API | 17:00 |
| winston-d | OK | 17:00 |
| e0ne | winston-d: I'm going to implement attach/detach too. it was already done for brickclient PoC | 17:00 |
| *** martyturner has quit IRC | 17:00 | |
| *** haomaiwang has quit IRC | 17:01 | |
| winston-d | so brickclient is a new project? | 17:01 |
| *** links has quit IRC | 17:01 | |
| *** haomaiwang has joined #openstack-cinder | 17:01 | |
| e0ne | winston-d: we decided to not introduce new project and move all to cinderclient | 17:02 |
| e0ne | winston-d: so I need to move code from my brickclient repo to cinderclient | 17:02 |
| DuncanT | smcginnis: Need to understand the implications - if it can be used to access other tenant data or something then it is a security bug | 17:02 |
| winston-d | ok, sounds good | 17:02 |
| wanghao | smcginnis: So you and John didn't agree to add a new column to DB, right? | 17:02 |
| smcginnis | wanghao: No, I think we were in agreement, but jgriffith may have more input on that. | 17:03 |
| *** merooney has quit IRC | 17:03 | |
| smcginnis | DuncanT: Can you explain that a little more? | 17:03 |
| DuncanT | e0ne: How do you intend to avoid confusion by having anything that says 'attach' in cinder client? | 17:03 |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Implement snapshots-related features for Block Device Driver https://review.openstack.org/253111 | 17:04 |
| e0ne | DuncanT: I would like to have something like local-attach | 17:04 |
| *** jgregor has quit IRC | 17:04 | |
| *** jgregor has joined #openstack-cinder | 17:05 | |
| e0ne | DuncanT: TBH, I've implemented it only for CLI, I'm not sure yet how to provide python API as a cinderclient extension | 17:05 |
| DuncanT | e0ne: I hate it. People *will* get confused | 17:05 |
| jgriffith | DuncanT: e0ne the local-attach seemed reasonable to me :( | 17:05 |
| e0ne | DuncanT: it should be something like cinderclient.brick.local_attach | 17:05 |
| scottda | e0ne: Will there also be local-terminate_connection or will that be wrapped in local-attach? | 17:05 |
| DuncanT | e0ne: Somebody will file a bug trying to use it inside a guest or something | 17:05 |
| e0ne | scottda: yes | 17:05 |
| scottda | ok, cool. I put a comment to that effect in the spec. | 17:06 |
| *** alonma has joined #openstack-cinder | 17:06 | |
| e0ne | scottda: I've added it to the spec | 17:06 |
| scottda | thx | 17:06 |
| *** yuriy_n17_ has joined #openstack-cinder | 17:06 | |
| *** martyturner has joined #openstack-cinder | 17:06 | |
| *** yuriy_n17_ has quit IRC | 17:07 | |
| e0ne | DuncanT: what do you mean? attach inside a guest is one of expected use cases | 17:07 |
| DuncanT | smcginess: ldev is in the volume_metadata - looks like it might be the provider_id - if changing it in the metadata allows you to change what backend volume your cidner vol points to, then that's a security bug | 17:07 |
| DuncanT | e0ne: For normal clouds, the guest can't access the cidner storage network | 17:07 |
| winston-d | e0ne: well, yes and no. For ironic, yes, but not for VM (especially VM with overlay network) | 17:07 |
| openstackgerrit | Mitsuhiro Tanino proposed openstack/cinder: Add validation for volume_type at copy_volume_to_image() https://review.openstack.org/254414 | 17:07 |
| jgriffith | DuncanT: e0ne I'm still unclear why you wouldn't explicity disable it by default | 17:07 |
| e0ne | DuncanT: good point, agree | 17:07 |
| *** diogogmt has joined #openstack-cinder | 17:08 | |
| e0ne | jgriffith: it will | 17:08 |
| jgriffith | but ok | 17:08 |
| e0ne | jgriffith: I won't add os-brick as a mandatory dependency | 17:08 |
| jgriffith | so disable by default, have a help that says what it's for and be done I thougt | 17:08 |
| *** thangp has quit IRC | 17:08 | |
| e0ne | jgriffith: so, by default, it won't work | 17:08 |
| jgriffith | e0ne: well, I meant a step further like something in policy that just completely disables it unless set explicitly | 17:09 |
| jgriffith | but it sounds like there's concerns with the naming itself etc.... | 17:09 |
| e0ne | jgriffith: I remember that point in our conversation | 17:09 |
| wanghao | smcginnis:jgriffith It's more easier to handle the string by adding a new column IMHO, but I'm fine if we want to use some existing column to store this replica volume provider id. | 17:09 |
| smcginnis | DuncanT: Ah, yeah. Definitely badness. | 17:09 |
| smcginnis | I would consider that a security bug. | 17:09 |
| jgriffith | seems like there's a group that wants a "new" client and a group (like me) that wants it in existing client | 17:09 |
| hemna | ildikov, ping | 17:10 |
| e0ne | jgriffith: need to do move investigation what is the best way to implement it | 17:10 |
| jgriffith | if we can't agree then I guess the "new client" group wins because it's silly to not do this | 17:10 |
| smcginnis | bbl | 17:10 |
| e0ne | jgriffith: and than't why my spec is not merged yet:( | 17:10 |
| jgriffith | wanghao: so I would prefer using an existing column if possible | 17:10 |
| *** alonma has quit IRC | 17:10 | |
| jgriffith | wanghao: if nothing else, maybe we could try that approach... if we decide we need to later we can revisit a new column? | 17:11 |
| winston-d | e0ne: do you have spec review link handy? | 17:11 |
| *** jungleboyj has quit IRC | 17:11 | |
| DuncanT | jgriffith: I can't disable something in a client | 17:11 |
| e0ne | winston-d: https://review.openstack.org/#/c/224124/13 | 17:11 |
| jgriffith | DuncanT: I think you can actually | 17:11 |
| *** rcernin has quit IRC | 17:11 | |
| e0ne | DuncanT: we can to not install os-brick | 17:11 |
| DuncanT | jgriffith: apt-get isntall python-cinderclient; cinder help; see something about attach. Try it. Get error. File bug. | 17:11 |
| winston-d | e0ne: thx | 17:12 |
| hemna | e0ne, so the client patch doesn't include a requirements.txt update to add os-brick | 17:12 |
| jgriffith | DuncanT: well, I'm saying I think I can hide the command altogether unless it's "turned on" | 17:12 |
| e0ne | DuncanT: it won't be an error. There will be a user-friendly message how to use it and what lilitations are | 17:12 |
| hemna | e0ne, is that intentional ? | 17:12 |
| wanghao | jgriffith: emm, it's fine to me. I will try it. | 17:12 |
| jgriffith | DuncanT: I'll see if I can code something up. Maybe I'm wrong, but I think I've done that before | 17:13 |
| DuncanT | jgriffith: Ok | 17:13 |
| e0ne | hemna: yes, because it's an extension | 17:13 |
| jgriffith | DuncanT: if I'm mistaken then fair enough | 17:13 |
| *** diogogmt_ has joined #openstack-cinder | 17:13 | |
| hemna | e0ne, so how does a user of the client know they have to install it ? | 17:13 |
| *** diogogmt has quit IRC | 17:13 | |
| *** diogogmt_ is now known as diogogmt | 17:13 | |
| hemna | run it, it fails | 17:13 |
| DuncanT | jgriffith: Sure, I might be missing something too | 17:13 |
| e0ne | hemna: cinder --help? | 17:13 |
| jgriffith | hemna: ok... this is kinda cra-cra | 17:14 |
| hemna | I'm just asking | 17:14 |
| hemna | trying to understand the use case | 17:14 |
| e0ne | hemna: if we want to add it as a mandatory requirement - we can discuss it later, it's not a big deal for us and package maintainers | 17:14 |
| wanghao | Well, it's too later in my TZ, so bye everyone~ | 17:14 |
| jgriffith | hemna: sorry... what I mean is.... bolting on os-brick as a req for cinderclient for one specific set of use cases | 17:14 |
| winston-d | wanghao: come on, it's only 1:14 am | 17:15 |
| hemna | e0ne, cinder --help poops out a boatload of stuff, I'm sure it would easily be missed | 17:15 |
| e0ne | hemna: it requires 'sudo permissions', so I think it's ok that the feature is disabled be default | 17:15 |
| jgriffith | hemna: let me see if I can hide things first and come back to this discussion | 17:15 |
| wanghao | winston-d: haha, | 17:15 |
| hemna | jgriffith, I'm not advocating adding it to requirements.txt | 17:15 |
| winston-d | wanghao: :) gdnite | 17:15 |
| hemna | I'm just trying to figure out what an end user's experience will be like | 17:15 |
| *** vgridnev has joined #openstack-cinder | 17:15 | |
| jgriffith | hemna: ahh... ok. Wasn't sure | 17:15 |
| hemna | and see if it makes sense and what we can do to make it better | 17:15 |
| hemna | if at all | 17:15 |
| hemna | cinder --help is almost useless IMHO | 17:16 |
| jgriffith | hemna: cool, yeah that's awesome | 17:16 |
| wanghao | winston-d: I need to get up early in 6:00 AM.... | 17:16 |
| e0ne | hemna: https://review.openstack.org/#/c/254878/1/cinderclient/v2/contrib/brick/brick/utils.py - line #19 | 17:16 |
| wanghao | winston-d: Good night too. | 17:16 |
| hemna | e0ne, so we do a try import somewhere else and dump out that error message ? | 17:17 |
| jgriffith | BTW... we have thing slike backup-create in their now :( | 17:17 |
| e0ne | hemna: yes | 17:17 |
| e0ne | hemna: https://review.openstack.org/#/c/254878/1/cinderclient/v2/contrib/brick/brick/client.py - line #29 | 17:17 |
| jgriffith | which has the same problem people are concerned about WRT confusion if things aren't enabled | 17:17 |
| hemna | ah yah I just saw that | 17:17 |
| hemna | ok I'll pull the patch and see what it says | 17:17 |
| hemna | maybe include the actual command to install it ? | 17:18 |
| *** leeantho has joined #openstack-cinder | 17:18 | |
| hemna | dunno | 17:18 |
| e0ne | hemna: it's WIP code now. I just try to implement an extension | 17:18 |
| hemna | no worries | 17:18 |
| e0ne | hemna: create better UX is on my plan | 17:18 |
| hemna | I'm not trying to criticize, I just want to understand the work flow and use case for end users. | 17:18 |
| *** jistr has quit IRC | 17:19 | |
| e0ne | hemna: got it, any feedback is welcome | 17:19 |
| hemna | e0ne, thanks for taking this on :) | 17:19 |
| hemna | my plate is way over full right now | 17:19 |
| *** ianbrown has quit IRC | 17:20 | |
| *** Apoorva has joined #openstack-cinder | 17:21 | |
| *** fthiagogv has joined #openstack-cinder | 17:21 | |
| *** electrocucaracha has joined #openstack-cinder | 17:21 | |
| openstackgerrit | Mikhail Khodos proposed openstack/cinder: Re-add Nexenta driver https://review.openstack.org/188877 | 17:22 |
| *** ianbrown has joined #openstack-cinder | 17:23 | |
| DuncanT | smcginnis: Looks like it isn't a security bug the ldev is shoved in the metadata but never used from there AFAICT - they put it in the provider_location as well, and use it from there | 17:23 |
| e0ne | hemna: thanks for help. | 17:23 |
| e0ne | hemna: http://paste.openstack.org/show/481345/ | 17:23 |
| *** fthiagogv has quit IRC | 17:23 | |
| hemna | ok cool | 17:23 |
| *** fthiagogv has joined #openstack-cinder | 17:23 | |
| hemna | we can craft a better message later before the patch lands, but that's a good start thanks. | 17:24 |
| e0ne | sure | 17:24 |
| electrocucaracha | hi there, We were playing around with horizon and we noticed that displaying bootable volume options is not working for non-admin users, from horizon source code perspective it's using same source code for all users. I'm wondering if this is a bug of Cinder server | 17:24 |
| *** dims has quit IRC | 17:24 | |
| e0ne | hemna, DuncanT: so, are you OK to make it inside cinderclient? | 17:25 |
| *** smoriya has quit IRC | 17:25 | |
| *** jaypipes has quit IRC | 17:25 | |
| hemna | e0ne, maybe explain what the command line call looks like for issuing the attach | 17:25 |
| hemna | and detach | 17:26 |
| DuncanT | e0ne: I don't understand what is being proposed at the moment. | 17:26 |
| e0ne | hemna: https://review.openstack.org/#/c/224124/13/specs/mitaka/use-cinder-without-nova.rst - line #64 | 17:26 |
| e0ne | DuncanT: please, take a look on spec ^^ | 17:26 |
| hemna | e0ne, cool thanks | 17:27 |
| DuncanT | e0ne: No, I hate it. I am entirely certain that somebody will try to run that from inside a guest and complain when it doesn't work | 17:27 |
| DuncanT | e0ne: I think it adds unnecessary confusion for a niche usecase | 17:28 |
| *** esker has joined #openstack-cinder | 17:28 | |
| *** isaacb has quit IRC | 17:28 | |
| DuncanT | e0ne: I think the cinder dev team underestimate how easily confused and damn hard work normal tenants can be | 17:28 |
| *** jordanP has quit IRC | 17:29 | |
| DuncanT | e0ne: We keep talking about needing smarter admins, who in real life often aren't there, tenants really do sit there poking likely looking buttons with no understanding of what they're doing. I've got a bug tracking system full of tickets to that effect | 17:29 |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/python-cinderclient: WIP. Add Python and CLI interfaces for os-brick as an extention https://review.openstack.org/254878 | 17:29 |
| *** salv-orlando has quit IRC | 17:30 | |
| *** thangp has joined #openstack-cinder | 17:30 | |
| winston-d | DuncanT: as an operator, I share your pain. | 17:30 |
| e0ne | DuncanT: we can make it hidden by default | 17:30 |
| DuncanT | e0ne: IMO bare metal attach has no place in the 'cinder' command line | 17:31 |
| hemna | DuncanT, the same can be said for a new client. | 17:31 |
| e0ne | DuncanT: I understand you, I'm asked about cinderclient.attach method almost once per month | 17:31 |
| e0ne | hemna: +1 | 17:31 |
| hemna | they'll pull down said brickclient and try running attach from inside the vm | 17:31 |
| hemna | and it'll fail | 17:31 |
| hemna | if they went through the trouble of pulling down cinderclient inside their guest, they'd do the same thing for brickclient | 17:32 |
| DuncanT | hemna: At least there you can say 'don't use brickclient, use cinder/nova, brick is for something different' | 17:32 |
| ildikov | hemna: hi, sorry I was on a call | 17:32 |
| hemna | users are stupid, they'll do dumb things | 17:32 |
| hemna | they shouldn't pull down cinderclient inside the guest either | 17:32 |
| hemna | it's really the same thing IMHO | 17:32 |
| DuncanT | hemna: 'cinder-client' looks like the right thing to be using to do stuff with cinder volumes.... 'brick' doesn't | 17:32 |
| hemna | DuncanT, which is the argument for putting the local-attach in cinderclient :P | 17:33 |
| DuncanT | hemna: I pull down cinder-client in guests all the time, e.g. guests crontab their own backups | 17:33 |
| hemna | ildikov, quick question. I want to test out the multiattach for nova side, so I can try and reproduce the bug https://bugs.launchpad.net/cinder/+bug/1471278 | 17:33 |
| openstack | Launchpad bug 1471278 in Cinder "target removal on detaching volume with multi-attach flag" [Undecided,New] - Assigned to Walt Boring (walter-boring) | 17:33 |
| DuncanT | hemna: I'm happy to make life harder for bare metal users to make it easier for 'normal' users | 17:33 |
| hemna | ildikov, which nova patch should I pull to get all of it ? | 17:33 |
| hemna | DuncanT, I'm just saying it's the same thing really. | 17:34 |
| hemna | it'll fail just the same either way inside the guest. | 17:34 |
| DuncanT | hemna: It is far easier to tell somebody 'brick-client is the wrong thing' than 'cinder local-attach is wrong, but cinder <everything else>' is right | 17:35 |
| *** egonzalez has quit IRC | 17:35 | |
| ildikov | hemna: theoretically this one: https://review.openstack.org/#/c/193135/ | 17:35 |
| hemna | DuncanT, I do see your point and understand. it's sucky either way really. | 17:35 |
| hemna | ildikov, ok I'll try that today and see how it goes :) | 17:36 |
| hemna | I'd like to get the lvm driver fixed, since it is our reference driver. | 17:36 |
| ildikov | hemna: although I have to admit that the tempest tests are still not green for it and I don't know what the exact issue is at the moment... :S | 17:36 |
| *** bardia has quit IRC | 17:36 | |
| DuncanT | hemna: yup | 17:36 |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Attach/detach volumes without Nova https://review.openstack.org/224124 | 17:36 |
| hemna | ildikov, who needs tempest anyway :) | 17:36 |
| ildikov | hemna: I took over the work and did a quick rebase, but then many things went wrong | 17:37 |
| ildikov | hemna: lol :) | 17:37 |
| *** p0rtal has quit IRC | 17:37 | |
| hemna | ildikov, thanks for the info. I'll try and get it working today. the patch looks like it won't induce merge conflicts, so hopefully the real code works :) | 17:37 |
| ildikov | hemna: I tried to set up tempest locally, but I think I will go for Devstack instead to see for myself what goes wrong and where | 17:38 |
| hemna | ok off to a beating....bbiab. | 17:38 |
| winston-d | if we have brick-client, what use cases would it serve? local-attach/detach for one. | 17:38 |
| *** hemna is now known as hemnafk | 17:38 | |
| ildikov | hemnafk: thanks for playing wiht it, let me know if you found any issues | 17:38 |
| *** sgotliv_ has quit IRC | 17:39 | |
| e0ne | winston-d: yes, only local attach/detach | 17:39 |
| *** gouthamr has quit IRC | 17:40 | |
| winston-d | nothing else? | 17:40 |
| e0ne | winston-d: only these two and get-connector | 17:41 |
| e0ne | winston-d: nothing else in the nearlies future | 17:41 |
| e0ne | DuncanT: each option (cinderclient or brickclient) will not work for everybody. we'll have confused users in any case | 17:42 |
| *** zhipeng has quit IRC | 17:42 | |
| *** timcl has joined #openstack-cinder | 17:43 | |
| *** changbl has quit IRC | 17:43 | |
| *** merooney has joined #openstack-cinder | 17:44 | |
| *** gouthamr has joined #openstack-cinder | 17:45 | |
| *** zhipeng has joined #openstack-cinder | 17:47 | |
| *** ndipanov has quit IRC | 17:47 | |
| *** salv-orlando has joined #openstack-cinder | 17:49 | |
| DuncanT | e0ne: I think I'm going to have a much easier job telling cloud users not to touch brick-client than trying to tell them that only some bits of cinder-client apply to them | 17:50 |
| *** sgotliv_ has joined #openstack-cinder | 17:51 | |
| *** salv-orlando has quit IRC | 17:52 | |
| *** martyturner has quit IRC | 17:55 | |
| *** lpetrut has quit IRC | 17:56 | |
| e0ne | DuncanT: the 3rd option: make it as a extension to cinderclient and install like 'pip install python-cinderclient && pip install python-cinderclient-brick' | 17:56 |
| e0ne | DuncanT: we just need to extend our cinderclient extensions API a bit for such case | 17:56 |
| e0ne | DuncanT: and your users won't see any local-attachment stuff for cinderclient w/o installing additional packages | 17:57 |
| e0ne | DuncanT: is it good for you? | 17:57 |
| DuncanT | e0ne: That sounds reasonable, yes | 17:58 |
| * e0ne changing the spec | 17:59 | |
| *** merooney has quit IRC | 17:59 | |
| *** hemnafk is now known as hemna | 17:59 | |
| *** markus_z has quit IRC | 17:59 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder-specs: Attach/detach volumes without Nova https://review.openstack.org/224124 | 18:00 |
| e0ne | DuncanT: please, review ^^ | 18:00 |
| *** haomaiwang has quit IRC | 18:01 | |
| *** haomaiwang has joined #openstack-cinder | 18:01 | |
| e0ne | hemna: ^^ | 18:01 |
| *** changbl has joined #openstack-cinder | 18:02 | |
| hemna | so python-cinderclient-brick ? | 18:03 |
| hemna | which is a post install extension to cinderclient ? | 18:03 |
| *** merooney has joined #openstack-cinder | 18:03 | |
| *** jkremer has joined #openstack-cinder | 18:04 | |
| *** jkremer has quit IRC | 18:04 | |
| *** rushiagr is now known as rushiagr_away | 18:05 | |
| *** jkremer has joined #openstack-cinder | 18:05 | |
| hemna | ildikov, ping | 18:06 |
| jgriffith | DuncanT: e0ne here's a quick and dirty example: https://gist.github.com/60515b3844238f102c49 | 18:06 |
| openstackgerrit | Merged openstack/cinder: Python 3: fix a lot of tests https://review.openstack.org/254198 | 18:06 |
| *** alonma has joined #openstack-cinder | 18:06 | |
| jgriffith | DuncanT: e0ne so that will get us "part" way there... that hides the commands unless --local is set | 18:06 |
| smcginnis | hemna: python-cinderclient-onlyuse-ifyouknow-whatyourdoing | 18:06 |
| jgriffith | DuncanT: e0ne Then on the Cinder API side you could enforce a policy to enable/disable it | 18:07 |
| *** rushiagr_away is now known as rushiagr | 18:07 | |
| jgriffith | DuncanT: e0ne in which case if a user enabled it in the client, calls it and they're not supposed to it will respond with "policy doesn't allow" | 18:07 |
| DuncanT | jgriffith: That's no worse than the two-part install I guess | 18:07 |
| jgriffith | DuncanT: e0ne I think there's much more that can be done to improve it | 18:08 |
| hemna | I like the idea of a pip install cinderclient-brick extension | 18:08 |
| jgriffith | hemna: that works too.. this is just "how" cinderclient would pull it in | 18:08 |
| hemna | yah cool | 18:08 |
| *** fthiagogv has quit IRC | 18:08 | |
| *** eduardo___ has quit IRC | 18:08 | |
| *** fthiagogv has joined #openstack-cinder | 18:09 | |
| *** sgotliv_ has quit IRC | 18:09 | |
| * DuncanT is ok with that idea, eitehr way | 18:09 | |
| *** timcl has left #openstack-cinder | 18:09 | |
| ildikov | hemna: pong | 18:09 |
| hemna | ildikov, so I'm getting a failure on a 2nd attach call | 18:10 |
| ildikov | the first is ok? | 18:10 |
| hemna | http://paste.openstack.org/show/481351/ | 18:10 |
| hemna | n-api is raising that | 18:10 |
| hemna | first attach is fine | 18:10 |
| *** alonma has quit IRC | 18:10 | |
| hemna | 2nd attach fails due to 'in-use' set. that shouldn't happen if the volume is multiattachable | 18:11 |
| *** rs007_ has quit IRC | 18:11 | |
| ildikov | hmm, that's not good | 18:12 |
| ildikov | when we stopped the work on the patches during summer the attach worked :( | 18:12 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/253753 | 18:12 |
| *** gouthamr has quit IRC | 18:12 | |
| ildikov | or my colleague lied... | 18:13 |
| hemna | http://paste.openstack.org/show/481352/ | 18:13 |
| hemna | gah | 18:13 |
| hemna | man pastebin is less than ideal for reading dumps | 18:13 |
| hemna | heh | 18:13 |
| hemna | anyway the volume dict has the "multiattach": true in it. | 18:13 |
| e0ne | jgriffith, DuncanT, hemna, smcginnis: team, please, lets make a decision and go forward with it. we are trying to get it landed since Vancouver summit :( | 18:13 |
| smcginnis | e0ne: +1 | 18:14 |
| jgriffith | e0ne: my vote is just implement it in a new module like I have in the example | 18:14 |
| jgriffith | e0ne: but I'll defer to the HP folks | 18:14 |
| jgriffith | e0ne: if we agree I can clean up that code and submit at least the foundation for it | 18:15 |
| e0ne | I'm ready to implement any solution which community will accept | 18:15 |
| hemna | honestly, I don't really care either way, just as long as users have the ability to do it. | 18:15 |
| e0ne | hemna: +1 | 18:15 |
| *** fthiagogv has quit IRC | 18:15 | |
| smcginnis | Agreed | 18:15 |
| *** fthiagogv has joined #openstack-cinder | 18:15 | |
| jgriffith | smcginnis: hemna DuncanT e0ne ok... so does that mean "go forward with the concept proposed"? | 18:16 |
| *** lcurtis has joined #openstack-cinder | 18:16 | |
| hemna | sure | 18:16 |
| jgriffith | #vote | 18:16 |
| jgriffith | +1 | 18:16 |
| *** martyturner has joined #openstack-cinder | 18:16 | |
| hemna | +1 | 18:16 |
| smcginnis | jgriffith: If DuncanT doesn't totally hate it anymore. | 18:16 |
| smcginnis | +1 | 18:16 |
| * e0ne can't +1 on own spec but | 18:16 | |
| e0ne | +1 | 18:16 |
| smcginnis | e0ne: Sure you can. :) | 18:16 |
| jgriffith | e0ne: :) | 18:16 |
| smcginnis | I see folks doing it all the time. Hah! | 18:16 |
| e0ne | set cheatmode=On | 18:17 |
| DuncanT | +1 | 18:17 |
| jgriffith | e0ne: I'm talking voting on actually using the cinderclient hack I just proposed in conjunction with policy | 18:17 |
| *** alonma has joined #openstack-cinder | 18:17 | |
| smcginnis | DuncanT: That works for you? | 18:17 |
| e0ne | +1 on any. both options looks good | 18:17 |
| smcginnis | e0ne: Yeah, the problem is I can see valid arguments on both sides. | 18:17 |
| DuncanT | As long as if I do 'pip install python-cinderclient; cinder help' and don't see anything about attach without doing some other step, I'm passibly happy | 18:18 |
| smcginnis | I don't want that to stall things though. | 18:18 |
| jgriffith | well now I'm confused but ok... whatevers; I'm going to move along now | 18:18 |
| *** markvoelker has quit IRC | 18:18 | |
| *** fthiagogv has quit IRC | 18:18 | |
| DuncanT | So I'm +1 on either proposal | 18:18 |
| smcginnis | DuncanT: Love it - passibly happy. :) | 18:18 |
| hemna | ildikov, crap. I just touched the nova/volume/cinder.py to add a debug log output and tried to bounce n-api, n-cpu and they both puked on startup | 18:18 |
| *** fthiagogv has joined #openstack-cinder | 18:18 | |
| *** rushiagr is now known as rushiagr_away | 18:18 | |
| DuncanT | jgriffith's or just the split install | 18:18 |
| hemna | ildikov, IncompatibleObjectVersion Version 1.18 of BlockDeviceMappingList is not supported | 18:18 |
| hemna | :( | 18:18 |
| jgriffith | DuncanT: ahhh | 18:18 |
| ildikov | hemna: bah :( | 18:18 |
| *** bardia has joined #openstack-cinder | 18:18 | |
| jgriffith | ok, well there's some options whatever works best for e0ne and everyone else | 18:19 |
| hemna | ildikov, *sigh* trying to debug this to figure it out. nova is always uber HARD | 18:19 |
| hemna | :( | 18:19 |
| ildikov | hemna: the version is bumped in the second patch in the chain | 18:19 |
| hemna | ildikov, guess I'll restack and see what happens. | 18:19 |
| ildikov | hemna: it's the obj: ... one IIRC | 18:20 |
| hemna | oh! | 18:20 |
| e0ne | 'pip install cinderclient-brick' is a good compromise to make everyone happy, IMO | 18:20 |
| hemna | bah | 18:20 |
| hemna | I'm an idiot. | 18:20 |
| hemna | heh. I stacked, then pulled the patch. | 18:20 |
| hemna | ok this should work, restacking. | 18:20 |
| ildikov | :) | 18:20 |
| e0ne | jgriffith: I like your proposal too but we need to confirm that DuncanT is OK for it too | 18:20 |
| *** abhi has joined #openstack-cinder | 18:20 | |
| DuncanT | e0ne: YEs | 18:21 |
| e0ne | DuncanT: to be clear: separate install (pip install cinderclient-brick) or John's proposal (cinder --local) | 18:21 |
| *** markvoelker has joined #openstack-cinder | 18:21 | |
| *** alonma has quit IRC | 18:22 | |
| DuncanT | e0ne: I'm fine with either. I guess John's proposal is easier | 18:22 |
| * e0ne modifies spec again | 18:22 | |
| jgriffith | e0ne: and I don't care much either... just so long as the pip install is just installing extensions and not a full-blown client | 18:22 |
| jgriffith | e0ne: they each have pros/cons neither a clear winner from what I can see | 18:23 |
| hemna | jgriffith, +1 | 18:23 |
| e0ne | jgriffith: +1 | 18:23 |
| DuncanT | I'm heading off, but I'll happily +1 either version in a spec | 18:23 |
| jgriffith | I wouldn't be surprised if folks like Dean Troyer or guitarzan might even already have cases where they do this sort of thing.. but I dunno | 18:24 |
| hemna | jgriffith, might be worth asking guitarzan about this as well | 18:24 |
| guitarzan | jgriffith: a cinderclient extension? some of the guys use one for this (detach) purpose right now | 18:26 |
| * guitarzan didn't read all the backlog | 18:26 | |
| e0ne | https://review.openstack.org/#/c/224124/ is updated with '-L/--local' option | 18:26 |
| guitarzan | for the record, we use a separate package for the extension | 18:27 |
| guitarzan | (I guess that's obvious) | 18:27 |
| jgriffith | guitarzan: yeah, we were talking between a seperate installable package for the extensions or just backing them directly in | 18:27 |
| jgriffith | guitarzan: sounds to me like the external package that gets imported is the "norm" | 18:28 |
| jgriffith | guitarzan: e0ne hemna You could use a combination of the approaches as well... | 18:28 |
| jgriffith | that way nobody would need "two" clients on their system to keep from getting confused :) | 18:28 |
| e0ne | jgriffith: I don't see any banefits on getting both options together | 18:29 |
| e0ne | 'pip install cinderclient-some-extension' should enable it by default. | 18:29 |
| e0ne | or we'll have a lot of questions like: I installed an extension, but it dosn't work | 18:30 |
| jgriffith | e0ne: KO | 18:30 |
| jgriffith | do the pip install extensions thing then | 18:30 |
| hemna | e0ne, +1 | 18:30 |
| hemna | jgriffith, +1 | 18:31 |
| jgriffith | just PLEASE for the love of all that is Beer, DO NOT CREATE A NEW CLIENT :) | 18:31 |
| * e0ne rolls back a spec | 18:31 | |
| hemna | :) | 18:31 |
| *** kfarr has quit IRC | 18:31 | |
| hemna | smcginnis, ping | 18:32 |
| bswartz | jgriffith: I read your ML post about snapshots and extend -- do you happen to know if there's tempest test coverage for that particular behavior (create from snapshots when volume was extended after snapshot)? | 18:32 |
| e0ne | the current state of https://review.openstack.org/#/c/224124/: propose to install python-cinder-brick as an extension | 18:32 |
| *** subhadeep_de has joined #openstack-cinder | 18:33 | |
| *** adrianofr has quit IRC | 18:33 | |
| *** adrianofr has joined #openstack-cinder | 18:33 | |
| *** p0rtal has joined #openstack-cinder | 18:33 | |
| smcginnis | hemna: sup | 18:33 |
| smcginnis | bswartz: DuncanT was going to look in to adding some tests. | 18:34 |
| *** angela-s has joined #openstack-cinder | 18:34 | |
| bswartz | smcginnis: +1000 | 18:34 |
| hemna | so this bug: https://bugs.launchpad.net/os-brick/+bug/1449492 | 18:34 |
| openstack | Launchpad bug 1449492 in os-brick "Cinder not working with IPv6 ISCSI" [Medium,In progress] - Assigned to Lukas Bezdicka (social-b) | 18:34 |
| hemna | it's not a bug in cinder, nor is it a bug in Nova. | 18:34 |
| hemna | brick is really doing the work | 18:34 |
| hemna | does it make sense to remove cinder and nova from that bug Affects list? | 18:34 |
| smcginnis | hemna: A bug in brick? | 18:35 |
| hemna | :P | 18:35 |
| smcginnis | hemna: Yeah, I think so. | 18:35 |
| smcginnis | :) | 18:35 |
| hemna | I'm trying to triage the os-brick bugs and make sure they are up to date. | 18:35 |
| smcginnis | hemna: Awesome! | 18:35 |
| hemna | ok so, in general, is it ok to remove cinder/nova affects for things like this | 18:35 |
| smcginnis | Yeah, I'd say remove it from cinder then if it's all inside brick. | 18:35 |
| hemna | folks that file the bugs don't know, and that's ok. | 18:35 |
| smcginnis | Yeah, I think so. | 18:35 |
| smcginnis | hemna: Yeah, agreed. Up to us I think to track down where the bug actually is. | 18:36 |
| hemna | ok cool. just wanted to make sure. | 18:36 |
| smcginnis | hemna: End users don't always know. | 18:36 |
| smcginnis | hemna: Thanks! | 18:36 |
| hemna | can't remove nova for some reason on that bug | 18:36 |
| hemna | maybe because there was a patch against nova originally | 18:37 |
| *** subhadeep_de has quit IRC | 18:37 | |
| smcginnis | hemna: Hmm, maybe. | 18:37 |
| smcginnis | Or maybe someone from nova needs to do it? | 18:38 |
| hemna | yah could be. | 18:38 |
| hemna | the os-brick patch is up and has 1 +2 | 18:38 |
| hemna | I'll just mark the nova side fixed once it's in. | 18:38 |
| smcginnis | Cool, that should work too. | 18:38 |
| e0ne | need to head home. will be back later tonight | 18:39 |
| smcginnis | e0ne: Thanks for sticking with this spec. | 18:40 |
| * e0ne hopes that spec will be approved today | 18:40 | |
| smcginnis | :) | 18:40 |
| e0ne | :) | 18:40 |
| *** jungleboyj has joined #openstack-cinder | 18:40 | |
| *** e0ne has quit IRC | 18:41 | |
| *** aix has quit IRC | 18:43 | |
| *** lpetrut has joined #openstack-cinder | 18:45 | |
| hemna | can you ssh into the cirros booted vms ? | 18:45 |
| hemna | what are the credentials ? | 18:46 |
| smcginnis | hemna: I thought so. It's been a while since I've used them. | 18:46 |
| guitarzan | hemna: it's something silly | 18:46 |
| hemna | I tried ssh cirros@<ip> no joy | 18:47 |
| guitarzan | there is a password | 18:47 |
| guitarzan | google is failing me | 18:48 |
| angela-s | cubswin:) | 18:48 |
| guitarzan | that's it | 18:48 |
| hemna | heh | 18:48 |
| hemna | which user ? | 18:48 |
| guitarzan | cirros | 18:48 |
| hemna | Permission denied, please try again. | 18:49 |
| hemna | bah | 18:49 |
| guitarzan | with the smiley? | 18:49 |
| hemna | oh! | 18:49 |
| angela-s | yah, i wasn't trying to be cute | 18:49 |
| angela-s | :) | 18:49 |
| hemna | lol | 18:49 |
| hemna | ok I missed that | 18:49 |
| hemna | thanks! | 18:49 |
| angela-s | haha | 18:50 |
| smcginnis | Hah, this could be a weird conversation for someone not familiar. | 18:50 |
| hemna | hehe | 18:50 |
| angela-s | when isn't the conversation weird on here? | 18:50 |
| hemna | ildikov, ok so I restacked, and the volume is attached to 2 vms, but after the 2nd attach the volume is still in 'attaching' state. | 18:51 |
| smcginnis | angela-s: Touche! | 18:51 |
| ildikov | hemna: hmm, that's interesting | 18:51 |
| hemna | trying to track it down | 18:52 |
| ildikov | hemna: I mean that state should be in-use after the first one, right? | 18:52 |
| ildikov | I'm just checking the code further, but to be honest the translations are putting me to hell | 18:53 |
| ildikov | the api code is duplicated in the sense of existing in both Cinder and Nova | 18:53 |
| hemna | ildikov, it is after the 1st attach | 18:53 |
| hemna | and the 2nd nova volume-attach call works | 18:53 |
| hemna | nova finds the volume and gives it to the 2nd vm | 18:54 |
| ildikov | and in Nova there is code for untranslate | 18:54 |
| hemna | but that's it. the volume should get moved from attaching back to in-use | 18:54 |
| ildikov | so the last step seems to be missing | 18:55 |
| hemna | yah | 18:55 |
| hemna | I'm trying to find out if cinder was called (attach) after nova is done | 18:55 |
| ildikov | but that should not depend on which attach you're in as far as I saw from the code | 18:56 |
| hemna | yah | 18:56 |
| ildikov | ... I mean it does not really touch the mechanism itself it just adds an 'or' to an 'if' to use the volume if it's 'in-use' too | 18:57 |
| ildikov | I couldn't even find whether it allows the re-use only if multiattach is true or always | 18:58 |
| hemna | check_attach is the only place nova really looks at that flag | 18:59 |
| hemna | afaik | 18:59 |
| ildikov | understanding Nova code is a full time job... | 18:59 |
| *** david-ly_ has joined #openstack-cinder | 19:00 | |
| hemna | hehe for sure | 19:00 |
| *** haomaiwang has quit IRC | 19:01 | |
| ildikov | hemna: yeah, right check_attach looks like the one | 19:01 |
| *** haomaiwang has joined #openstack-cinder | 19:01 | |
| *** david-ly_ is now known as david-lyle_ | 19:03 | |
| *** david-lyle has quit IRC | 19:03 | |
| *** electrocucaracha has quit IRC | 19:04 | |
| *** david-lyle_ is now known as david-lyle | 19:04 | |
| *** salv-orlando has joined #openstack-cinder | 19:05 | |
| hemna | ildikov, strange | 19:07 |
| *** akerr is now known as akerr_away | 19:07 | |
| hemna | I did another test, and nova never seems to call cinder after the 2nd attach completes | 19:07 |
| hemna | ildikov, http://paste.openstack.org/show/481362/ | 19:09 |
| hemna | there should have been another os-attach call to the c-api | 19:09 |
| *** Apoorva has quit IRC | 19:09 | |
| *** zhipeng has quit IRC | 19:10 | |
| *** geguileo has quit IRC | 19:10 | |
| hemna | yup, nova never calls attach after the 2nd one completes | 19:10 |
| ildikov | bah | 19:10 |
| *** zhipeng has joined #openstack-cinder | 19:11 | |
| hemna | http://paste.openstack.org/show/481363/ | 19:12 |
| *** electrocucaracha has joined #openstack-cinder | 19:13 | |
| ildikov | hemna: this patch modifies the compute part: https://review.openstack.org/#/c/193134/20 | 19:14 |
| ildikov | I cannot see anything that would affect calling out to Cinder | 19:15 |
| *** e0ne has joined #openstack-cinder | 19:15 | |
| hemna | ildikov, I think I found it..... | 19:16 |
| ildikov | shoot | 19:16 |
| hemna | https://github.com/openstack/nova/blob/master/nova/virt/block_device.py#L305 | 19:16 |
| hemna | it's only going to call cinder attach if the volume attach_status was detached | 19:17 |
| hemna | during the 2nd attach it will be 'attached' | 19:17 |
| *** alonma has joined #openstack-cinder | 19:17 | |
| hemna | so it won't call volume_api.attach | 19:17 |
| hemna | thus leaving the volume in attaching state | 19:18 |
| hemna | that's like 308 in the updated patchset | 19:18 |
| hemna | line | 19:18 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Move retype quota checks to API https://review.openstack.org/254980 | 19:18 |
| hemna | but that check is the same as line 305 on github | 19:18 |
| ildikov | hemna: https://review.openstack.org/#/c/193135/8/nova/virt/block_device.py | 19:19 |
| ildikov | hemna: resolving merge conflicts on a totally unknown code :( | 19:20 |
| hemna | ugh | 19:20 |
| ildikov | hemna: sorry for wasting your time on this | 19:20 |
| hemna | no worries | 19:20 |
| hemna | I'm glad I can help | 19:20 |
| ildikov | hemna: I triple checked it several times and there are still tiny things like this :.( | 19:21 |
| ildikov | by tiny I mean the size | 19:21 |
| hemna | yah that's one of the issues in nova. it's so amazingly complex | 19:21 |
| hemna | one small thing can break it all | 19:21 |
| ildikov | will upload a new version in a few mins | 19:22 |
| *** alonma has quit IRC | 19:22 | |
| hemna | ok thanks!! | 19:22 |
| *** Apoorva has joined #openstack-cinder | 19:24 | |
| openstackgerrit | Apoorva Deshpande proposed openstack/cinder: Retype functionality in Tintri driver https://review.openstack.org/255449 | 19:25 |
| *** sgotliv_ has joined #openstack-cinder | 19:26 | |
| *** wanghao has quit IRC | 19:27 | |
| openstackgerrit | Alex O'Rourke proposed openstack/cinder: 3PAR: Implement v2 replication (unmanaged) https://review.openstack.org/231245 | 19:27 |
| ildikov | hemna: I removed the check, new version is up | 19:27 |
| hemna | ok thanks. | 19:29 |
| hemna | ildikov, gerrit says merge conflict | 19:30 |
| hemna | fwiw | 19:30 |
| ildikov | I shouldn't have mentioned the joy of rebasing this code :) | 19:31 |
| hemna | wee! | 19:32 |
| *** p0rtal_ has joined #openstack-cinder | 19:32 | |
| *** subhadeep_de has joined #openstack-cinder | 19:32 | |
| hemna | worked! | 19:33 |
| hemna | after 2nd attach status = 'in-use' | 19:33 |
| hemna | phew! | 19:34 |
| e0ne | hemna: did you start any activities related to force detach? | 19:34 |
| hemna | e0ne, no | 19:34 |
| hemna | e0ne, it's on my list, but unfortunately nova is closed | 19:35 |
| e0ne | ok, so I can start work on spec | 19:35 |
| hemna | so, we are toast | 19:35 |
| e0ne | nova closed for new specs in mitaka? | 19:35 |
| ildikov | hemna: awesome, thanks :) | 19:35 |
| *** salv-orlando has quit IRC | 19:36 | |
| ildikov | hemna: the merge conflict was in a test, so the code is not worse than it was | 19:36 |
| hemna | :) | 19:36 |
| *** p0rtal has quit IRC | 19:36 | |
| ildikov | e0ne: yeap, they closed last week | 19:36 |
| e0ne | :( | 19:36 |
| hemna | :( | 19:37 |
| hemna | it seems like you get maybe 4 weeks to add nova-specs after the summit | 19:37 |
| hemna | and that's it. | 19:37 |
| e0ne | anyway, we can start with cinder side | 19:37 |
| hemna | e0ne, sure. | 19:37 |
| hemna | unused code though, which sux | 19:37 |
| hemna | anyway, | 19:37 |
| e0ne | and propose for N release in nova | 19:37 |
| ildikov | e0ne: you can still add one to the backlog folder | 19:37 |
| e0ne | hemna: +1 | 19:38 |
| ildikov | and then ping guys as soon as possible close to the end of Mitaka or very early N | 19:38 |
| e0ne | ildikov: thanks for the advise | 19:38 |
| ildikov | e0ne: np, I wish I would know any better option | 19:39 |
| ildikov | hemna: quick question | 19:39 |
| ildikov | do you happen to have n-api output too from your tests? | 19:40 |
| *** diablo_rojo has quit IRC | 19:40 | |
| hemna | yah | 19:41 |
| ildikov | I have a few failing tempest test cases, where the 'serverId' field in the attachments seems to be null | 19:42 |
| hemna | ildikov, http://paste.openstack.org/show/481365/ | 19:42 |
| hemna | jgriffith, ping | 19:42 |
| hemna | jgriffith, https://gist.github.com/WaltHP/9a4f4913ddd0012271b8 | 19:44 |
| hemna | Target 2,3,4 have multiple attaches to the same host | 19:44 |
| *** _cjones_ has joined #openstack-cinder | 19:45 | |
| openstackgerrit | Yuriy Nesenenko proposed openstack/cinder: Small refactoring in test_admin_actions.py https://review.openstack.org/253541 | 19:45 |
| ildikov | hemna: tnx | 19:45 |
| hemna | ildikov, anything else ? | 19:45 |
| ildikov | hemna: not for now, I will try to set up an env too to be able to test manually | 19:46 |
| *** briancurtin has quit IRC | 19:46 | |
| hemna | ok cool. | 19:46 |
| hemna | thanks so much for the help | 19:46 |
| ildikov | hemna: waiting for the gate and even the local test runs takes ages | 19:46 |
| *** merooney has quit IRC | 19:46 | |
| * ildikov thanks | 19:46 | |
| hemna | yah :( | 19:46 |
| hemna | it's a painful process for sure | 19:46 |
| *** martyturner has quit IRC | 19:46 | |
| *** briancurtin has joined #openstack-cinder | 19:47 | |
| ildikov | hemna: is there anything with the Devstack env I should keep an eye on, or just spin up an env and it should work? | 19:47 |
| hemna | I didn't do anything special | 19:47 |
| hemna | I just fired up another vagrant and pulled the nova patchset and stacked. | 19:48 |
| hemna | just remember to create the lvm volume with the --allow-multiattach flag | 19:48 |
| hemna | cinder create --allow-multiattach --name test 1 | 19:48 |
| hemna | after that and the updated nova patch, it worked. | 19:48 |
| hemna | the detach bug is still there with lvm fwiw | 19:48 |
| hemna | which is what I'm trying to figure out how to fix w/o disabling multiattach for it. | 19:49 |
| ildikov | hemna: ok, thanks for the advice and also for checking that bug | 19:49 |
| *** diablo_rojo has joined #openstack-cinder | 19:49 | |
| hemna | np | 19:50 |
| *** martyturner has joined #openstack-cinder | 19:50 | |
| *** electrocucaracha has quit IRC | 19:52 | |
| *** harlowja_at_home has quit IRC | 19:54 | |
| *** john5223 is now known as zz_john5223 | 19:56 | |
| *** fthiagogv has quit IRC | 19:57 | |
| *** david-lyle has quit IRC | 19:58 | |
| *** ociuhandu has quit IRC | 19:58 | |
| *** david-lyle has joined #openstack-cinder | 19:58 | |
| *** ChubYann has joined #openstack-cinder | 19:59 | |
| *** akerr_away is now known as akerr | 19:59 | |
| *** akerr is now known as akerr_away | 19:59 | |
| hemna | so....crap | 19:59 |
| *** akerr_away is now known as akerr | 19:59 | |
| hemna | nova never sends the host_name along in the attach call to cinder | 19:59 |
| e0ne | hm.... | 19:59 |
| hemna | if it did | 19:59 |
| *** mrnobody has joined #openstack-cinder | 20:00 | |
| *** merooney has joined #openstack-cinder | 20:00 | |
| *** haomaiwang has quit IRC | 20:01 | |
| hemna | we could test the attachments table | 20:01 |
| hemna | and track it that way | 20:01 |
| openstackgerrit | Mitsuhiro Tanino proposed openstack/cinder: Add validation for volume_type of volume object https://review.openstack.org/254414 | 20:01 |
| *** haomaiwang has joined #openstack-cinder | 20:01 | |
| e0ne | hemna: so host_name is not used anyware:( | 20:04 |
| hemna | it's easy to add | 20:05 |
| hemna | I hacked ildikov's patch and going to see if that gets it in the table | 20:05 |
| hemna | the lvm driver can use it to see if it has any more attachments to that host | 20:05 |
| hemna | and then use that to decide to remove the target or not. | 20:05 |
| *** hemna is now known as hemnalunch | 20:06 | |
| scottda | e0ne: We are working to get this Nova force-detach solution through: https://review.openstack.org/#/c/184537/9 | 20:09 |
| scottda | e0ne: This spec for a new API died after many months and much pain: https://review.openstack.org/#/c/84048/44 | 20:09 |
| *** salv-orlando has joined #openstack-cinder | 20:10 | |
| e0ne | scottda: thanks! I'll take a look on both patches tomorrow | 20:10 |
| scottda | e0ne: There's an ML thread about the contentious issues: https://review.openstack.org/#/c/84048/44 | 20:10 |
| scottda | e0ne: And it is to be discussed at Nova meeting this week; December 10th 2015 1400 UTC | 20:11 |
| e0ne | scottda: good, I'll be there | 20:12 |
| scottda | Let me know if you want to discuss, I've a long history here :) | 20:12 |
| *** alonma has joined #openstack-cinder | 20:12 | |
| e0ne | scottda: I have to read spec, patch and all comments before | 20:13 |
| scottda | You could start with the ML thread, might save some time.... | 20:13 |
| scottda | Read the patch for an idea of where this is likely to go... | 20:14 |
| scottda | Short summary of spec for API change is: folks thought Cinder API should be made more robust so this doesn't happen (I discussed doing this in the future releases at Tokyo summit)... | 20:15 |
| scottda | and people didn't want to change the Nova API with a --force option for volume detach | 20:15 |
| scottda | But there's a lot of back and forth | 20:15 |
| *** vgridnev has quit IRC | 20:17 | |
| *** xyang1 has quit IRC | 20:17 | |
| *** alonma has quit IRC | 20:17 | |
| jgriffith | hemnalunch: rigth, each with it's own target/nexus | 20:17 |
| *** xyang1 has joined #openstack-cinder | 20:17 | |
| e0ne | scottda: I need to read more to be know what nova need | 20:18 |
| *** xyang has joined #openstack-cinder | 20:18 | |
| e0ne | scottda: I don't want to start it at 10pm | 20:18 |
| jgriffith | hemnalunch: oh.. wait, what? | 20:19 |
| scottda | e0ne: Cool. The use case of cinder-attach-without-nova is interesting because we might find misbehaving client of that API will need force detach as well. It would be nice if we could do all the Cinder-side force-detach without any help from the consumer. | 20:20 |
| *** alonma has joined #openstack-cinder | 20:20 | |
| e0ne | scottda: yea, that's why I'm interesting on it | 20:20 |
| *** p0rtal has joined #openstack-cinder | 20:21 | |
| scottda | Cinder has force-detach API already, but it needs a connector passed in. | 20:22 |
| e0ne | scottda: we discussed at summit, that we want to save this info in the DB | 20:22 |
| scottda | yes. hemnalunch was going to work on that, but he might need some help. | 20:23 |
| scottda | I hope to have some time for this after the new year. | 20:23 |
| *** p0rtal_ has quit IRC | 20:24 | |
| *** alonma has quit IRC | 20:24 | |
| *** merooney has quit IRC | 20:26 | |
| *** alonma has joined #openstack-cinder | 20:26 | |
| *** mrnobody has quit IRC | 20:29 | |
| *** alonma has quit IRC | 20:30 | |
| *** Lee1092 has quit IRC | 20:31 | |
| *** alonma has joined #openstack-cinder | 20:32 | |
| *** esker has quit IRC | 20:33 | |
| *** alonma has quit IRC | 20:36 | |
| *** alonma has joined #openstack-cinder | 20:38 | |
| openstackgerrit | Simon Dodsley proposed openstack/cinder: Update Pure REST API supported version numbers in Pure driver https://review.openstack.org/255417 | 20:39 |
| *** belmoreira has joined #openstack-cinder | 20:41 | |
| *** alonma has quit IRC | 20:42 | |
| *** alonma has joined #openstack-cinder | 20:44 | |
| *** alonma has quit IRC | 20:48 | |
| *** xyang has quit IRC | 20:48 | |
| *** dims has joined #openstack-cinder | 20:49 | |
| *** alonma has joined #openstack-cinder | 20:50 | |
| *** merooney has joined #openstack-cinder | 20:53 | |
| *** alonma has quit IRC | 20:55 | |
| *** krtaylor has quit IRC | 20:56 | |
| *** Apoorva has quit IRC | 20:57 | |
| *** dims has quit IRC | 20:58 | |
| *** electrocucaracha has joined #openstack-cinder | 20:58 | |
| *** jseiler has quit IRC | 20:58 | |
| *** baumann1 has joined #openstack-cinder | 21:00 | |
| *** dims has joined #openstack-cinder | 21:00 | |
| *** haomaiwang has quit IRC | 21:01 | |
| *** dims has quit IRC | 21:01 | |
| *** 17WAA61QI has joined #openstack-cinder | 21:01 | |
| *** baumann has quit IRC | 21:02 | |
| *** xyang has joined #openstack-cinder | 21:04 | |
| *** akerr has quit IRC | 21:05 | |
| *** merooney has quit IRC | 21:05 | |
| *** raildo is now known as raildo-afk | 21:06 | |
| *** p0rtal_ has joined #openstack-cinder | 21:06 | |
| *** hemnalunch is now known as hemna | 21:08 | |
| hemna | back | 21:08 |
| openstackgerrit | John Griffith proposed openstack/cinder: Add cluster Pairing setup to SolidFire driver https://review.openstack.org/238235 | 21:09 |
| *** ctina_ has quit IRC | 21:09 | |
| *** lpetrut has quit IRC | 21:09 | |
| hemna | scottda, so yah the connector needs to get saved, but nova has to change to pass it at attach_volume time. | 21:10 |
| *** p0rtal has quit IRC | 21:10 | |
| scottda | oh yeah, that's right. Crap | 21:10 |
| jgriffith | xyang: ok, I *think* I caught all the things that got lost in the rebases :( | 21:10 |
| hemna | yup | 21:10 |
| scottda | hemna: IF that counts as a Nova API change, it needs a spec. Which means not until N release | 21:11 |
| scottda | We could still do force-detach-all in Mitaka. Unless I've forgotten something else... | 21:11 |
| openstackgerrit | Alex O'Rourke proposed openstack/cinder: 3PAR: Implement v2 replication (unmanaged) https://review.openstack.org/231245 | 21:12 |
| hemna | scottda, well the nova api itself isn't changing | 21:13 |
| hemna | but we need to update cinderclient to accept the connector and get nova to pass the connector at attach time. | 21:13 |
| scottda | yeah, true. Might be able to get this in Nova then. | 21:13 |
| e0ne | in such case, we need to release cinderclient very fast | 21:14 |
| scottda | We can still go forward with cinderclient and cinder server changes. | 21:14 |
| hemna | yup | 21:14 |
| scottda | e0ne: I thought you were going to sleep sometime :) | 21:14 |
| smcginnis | We should be able to get an update out within a couple days once we know it's ready. | 21:14 |
| smcginnis | scottda: Yeah, no kidding. e0ne's pulling an all nighter! | 21:15 |
| smcginnis | No sleep until you get your spec approved! :P | 21:15 |
| scottda | How about the cinder changes. Does that need a spec? If so.... | 21:15 |
| e0ne | what is 'to sleep'? | 21:15 |
| hemna | scottda, I think so. we need a new db column in the volume_attachment table | 21:15 |
| smcginnis | Cinder changes for multiattach? | 21:16 |
| hemna | to store the connector | 21:16 |
| hemna | and then an API change to accept it, etc | 21:16 |
| smcginnis | hemna: Oh, yeah. We probably should have a spec so it's at least documented. | 21:16 |
| scottda | smcginnis: cinder changes to store the volume connector at attach time, so we can use if for a 'force-detach' from cinder-only, without nova help | 21:16 |
| *** erlon has quit IRC | 21:16 | |
| scottda | amongst other reasons | 21:16 |
| smcginnis | Was originally thinking it would be just a bug fix, but that's probably significant enough we should do a spec. | 21:16 |
| hemna | yah | 21:16 |
| hemna | damn | 21:16 |
| smcginnis | :) | 21:17 |
| hemna | so passing host_name along with the attach call fails | 21:17 |
| hemna | cinder replies with 400 | 21:17 |
| scottda | hemna: I *think* I can work on this .....need to have my annual 5 minute conversation with my manager next week. | 21:17 |
| e0ne | we've got one issue whith this | 21:17 |
| e0ne | we can connector info for iscsi, nfs | 21:17 |
| hemna | FC | 21:17 |
| e0ne | but we can't do it in he same way for rbd | 21:18 |
| hemna | rbd doesn't add anything to the connector | 21:18 |
| e0ne | yes | 21:18 |
| hemna | as well as the other vendor connectors | 21:18 |
| e0ne | what about other protocols? I didn't check them | 21:18 |
| e0ne | ok, sounds good | 21:18 |
| hemna | https://github.com/openstack/os-brick/blob/master/os_brick/initiator/connector.py#L94 | 21:19 |
| hemna | the reply to that is the initiator connector dict | 21:19 |
| *** eharney has quit IRC | 21:19 | |
| *** thangp has quit IRC | 21:24 | |
| *** belmoreira has quit IRC | 21:26 | |
| *** lpetrut has joined #openstack-cinder | 21:27 | |
| e0ne | will try to sleep. it sound like a good thing:). bb tomorrow | 21:32 |
| *** lpetrut has quit IRC | 21:32 | |
| *** p0rtal_ has quit IRC | 21:32 | |
| *** e0ne has quit IRC | 21:32 | |
| *** ianbrown has quit IRC | 21:33 | |
| *** p0rtal has joined #openstack-cinder | 21:33 | |
| *** chris_morrell has quit IRC | 21:34 | |
| *** chris_morrell has joined #openstack-cinder | 21:35 | |
| *** p0rtal has quit IRC | 21:37 | |
| *** Apoorva has joined #openstack-cinder | 21:39 | |
| *** p0rtal has joined #openstack-cinder | 21:44 | |
| openstackgerrit | Alex O'Rourke proposed openstack/cinder: 3PAR: Implement v2 replication (unmanaged) https://review.openstack.org/231245 | 21:44 |
| *** zz_john5223 is now known as john5223 | 21:52 | |
| openstackgerrit | Alex O'Rourke proposed openstack/cinder: LeftHand: Implement v2 replication (unmanaged) https://review.openstack.org/255544 | 21:52 |
| *** nikeshm has quit IRC | 21:54 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/253753 | 21:55 |
| openstackgerrit | John Griffith proposed openstack/cinder: Add replicate by volume-type to SolidFire driver https://review.openstack.org/238276 | 21:55 |
| hemna | arg | 21:56 |
| *** 17WAA61QI has quit IRC | 22:01 | |
| *** garthb has joined #openstack-cinder | 22:01 | |
| *** haomaiwang has joined #openstack-cinder | 22:01 | |
| smcginnis | hemna: It's not talk like a pirate day, is it? :P | 22:02 |
| hemna | it should be | 22:02 |
| smcginnis | hemna: Missed it. It was Sept 19th: http://www.talklikeapirate.com/ | 22:03 |
| smcginnis | Hah | 22:03 |
| *** edmondsw has quit IRC | 22:04 | |
| *** electrocucaracha has quit IRC | 22:05 | |
| jgriffith | smcginnis: ye are too late | 22:07 |
| jgriffith | smcginnis: tis talketh liketh Shakespeare day | 22:08 |
| smcginnis | Surely you jest. It doth on current daylight breaks? | 22:08 |
| smcginnis | (Or something like that!) | 22:09 |
| jgriffith | smcginnis: well play'd sir! | 22:09 |
| smcginnis | Hah! | 22:09 |
| hemna | so... it seems the sqlalchemy Volume object doesn't have the volume attachments | 22:11 |
| hemna | https://github.com/openstack/cinder/blob/master/cinder/db/sqlalchemy/models.py#L119 | 22:12 |
| hemna | so when a driver gets called w/ the Volume model object the attachments aren't in there. | 22:12 |
| hemna | It looks like the consistencygroup var is being created in the Volume object | 22:12 |
| hemna | but I'm not sure how to get the attachments added to it automatically | 22:13 |
| hemna | https://github.com/openstack/cinder/blob/master/cinder/db/sqlalchemy/models.py#L178-L182 | 22:13 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Show qos_specs_id based on policy https://review.openstack.org/241659 | 22:13 |
| hemna | ^^ I presume that adds the consistency group object in the Volume object ? | 22:13 |
| DuncanT | To call or not to call, whether 'tis nobler in the code to accept anonymous attachments, or let them go unattached, and in doing so take battle against all previous versions? | 22:13 |
| hemna | :) | 22:14 |
| smcginnis | Now that is well played! ^^ | 22:14 |
| smcginnis | hemna: Isn't dulek our sqlalchemy expert? :) | 22:14 |
| hemna | If I can get that list of attachments added to the volume object, I think I have a solution for lvm | 22:15 |
| *** crose has joined #openstack-cinder | 22:15 | |
| hemna | which would be helpful for any driver that does the same thing that lvm does (1 export per host) | 22:15 |
| openstackgerrit | John Griffith proposed openstack/cinder: Add snapshot ID's to volume detail view https://review.openstack.org/255571 | 22:16 |
| smcginnis | hemna: Mine does. | 22:16 |
| jgriffith | DuncanT: BRAVO' FRIKIN'O | 22:16 |
| smcginnis | FYI: grenade tests are probably going to fail until this makes it through: https://review.openstack.org/#/c/253753/10 | 22:17 |
| jgriffith | DuncanT: err... "BravFrikin'O" | 22:17 |
| jgriffith | smcginnis: and my work here is done for the day... off to the bar | 22:17 |
| jgriffith | LOL | 22:17 |
| hemna | heh | 22:17 |
| smcginnis | Hehe, yep | 22:18 |
| mc_nair | hemna: I think you should be able to access the volume_attachments using volume.volume_attachment and it will lazy load it from the DB once you do that (https://github.com/openstack/cinder/blob/master/cinder/objects/volume.py#L295) | 22:18 |
| hemna | mc_nair, hrmm | 22:18 |
| hemna | I can try at least..... | 22:18 |
| openstackgerrit | Angela Smith proposed openstack/cinder: Adds HTTPS southbound connector for Brocade FC Zone Driver https://review.openstack.org/184799 | 22:18 |
| openstackgerrit | Nate Potter proposed openstack/cinder: Show qos_specs_id based on policy https://review.openstack.org/241659 | 22:19 |
| mc_nair | hemna: think that won't actually get loaded til you access it | 22:19 |
| mc_nair | *hopefully* | 22:19 |
| hemna | hah | 22:20 |
| hemna | what do you know | 22:20 |
| hemna | mc_nair, that worked | 22:20 |
| hemna | hrmm.... | 22:20 |
| mc_nair | hemna: uh oh- it worked and you're still hrmm-ing? | 22:20 |
| openstackgerrit | Simon Dodsley proposed openstack/cinder: Update Pure REST API supported version numbers in Pure driver https://review.openstack.org/255417 | 22:21 |
| hemna | mc_nair, it worked! | 22:21 |
| hemna | :) | 22:21 |
| hemna | I'm continuing my hatched job | 22:21 |
| *** dave-mccowan has quit IRC | 22:21 | |
| mc_nair | hemna: haha cool - just wanted to make sure something wasn't still off with this in particular. Good luck | 22:22 |
| mc_nair | hemna: btw, scrolled up and saw you saying you're plates pretty full, if there's some coding pieces you could use a hand with lemme know and I'd be happy to help out with them | 22:23 |
| DuncanT | Carving from the carcass a choice cut, a pound of fair code to wet the mouth and sate the appetites of the clammering hordes who scream for bloody and multiple attachments? | 22:23 |
| *** crose has quit IRC | 22:24 | |
| jgriffith | DuncanT: I must say I'm rather impressed... don't know what the hell you just said, but just the same :) | 22:24 |
| smcginnis | jgriffith: Is that any different than normal? | 22:24 |
| hemna | mc_nair, ok thanks man | 22:24 |
| jgriffith | smcginnis: that was my next line :) | 22:24 |
| smcginnis | :P | 22:25 |
| *** ebalduf has quit IRC | 22:25 | |
| *** alejandrito has quit IRC | 22:26 | |
| mc_nair | DuncanT: I'm sold. You have a past life as an Elizabethan salesman? | 22:27 |
| DuncanT | jgriffith: If a fool's capering is taken as an oracle by the wise who are ignored by the fool, then is not the fool wise and the wise fools? What we taketh in our sup is by our own hand to mouth poured and so no man may claim his fate in another's hands but that he choice that place that it was put and so guided it himself? | 22:27 |
| hemna | ok so terminate_connection has the connector, but the lvm driver nukes the target during remove_export | 22:27 |
| hemna | so close...yet so far | 22:27 |
| DuncanT | mc_nair: Ex theatre tech who'd bum beer between shows for such performances... I'm out of practice | 22:28 |
| scottda | hemna: Won't we want to put logic in (the manager) to only call remove_export upon the last attachment going away during call to detach() | 22:30 |
| scottda | Else the individual drivers will have to deal with it. | 22:30 |
| mc_nair | DuncanT: haha nice. Beer... the great motivator | 22:30 |
| hemna | could be | 22:30 |
| jgriffith | scottda: but that only works if you're attached on same host | 22:31 |
| hemna | I'm just hacking around at the moment | 22:31 |
| jgriffith | scottda: otherwise you should call it on the node | 22:31 |
| jgriffith | scottda: and by the way, multi-attach to the same node should just share the same connection anyway | 22:31 |
| scottda | jgriffith: Is that true for all drivers? | 22:31 |
| jgriffith | scottda: I suppose that they can do whatever they want, but I don't see a reason why not | 22:32 |
| jgriffith | scottda: make management easier that's for sure | 22:32 |
| scottda | I though 3par did it one way and Lefthand the other hemna ? | 22:32 |
| *** adrianofr has quit IRC | 22:33 | |
| jgriffith | scottda: my point was just that before rewriting everything for a specific backend might be good to look a little more at things | 22:33 |
| jgriffith | anyway... meeting | 22:33 |
| scottda | jgriffith: I agree. It would suck to be specific to the back end. | 22:33 |
| *** david-lyle has quit IRC | 22:34 | |
| hemna | http://paste.openstack.org/show/481399/ | 22:35 |
| hemna | ok | 22:35 |
| scottda | But if some drivers have 1 connection/host for multiple attaches, and some have 1 connection /attach that's a problem | 22:35 |
| hemna | it looks like that would 'work' | 22:35 |
| hemna | would have to migrate that logic into the manager | 22:35 |
| *** baumann1 has quit IRC | 22:35 | |
| hemna | instead of the hack I did in the lvm.py | 22:35 |
| scottda | ship it | 22:36 |
| hemna | 3par does a new export for every volume, even on the same host | 22:36 |
| hemna | lefthand acts like lvm | 22:36 |
| scottda | Does that logic look at the host field? | 22:36 |
| hemna | 1 export per host only | 22:36 |
| hemna | https://gist.github.com/WaltHP/7676c5c6d23c8b84e247 | 22:36 |
| hemna | my hack to lvm.py | 22:36 |
| scottda | i.e. can it be made generic for both cases (LH and 3par style) | 22:36 |
| hemna | scottda, ^^ | 22:36 |
| hemna | let me rework that hack and move the code into the manager from lvm.py | 22:37 |
| scottda | looks good | 22:38 |
| hemna | the good thing is, that the nova multiattach code hasn't landed yet | 22:38 |
| scottda | right, we can still influence that | 22:38 |
| hemna | and this required me changing nova code to pass the host | 22:38 |
| *** alonma has joined #openstack-cinder | 22:38 | |
| *** briancurtin has quit IRC | 22:38 | |
| scottda | Well, we might as well work to get whatever changes we want into Nova. Host, connector, kitchen sink | 22:39 |
| *** david-ly_ has joined #openstack-cinder | 22:39 | |
| hemna | hrmm | 22:39 |
| *** dilip has quit IRC | 22:39 | |
| hemna | who calls manager.remove_export | 22:39 |
| *** patrickeast has quit IRC | 22:40 | |
| hemna | ah yes our friend detach_volume | 22:40 |
| guitarzan | wait, did hemna just say it was good that nova's code hasn't landed? | 22:40 |
| hemna | hehe :) | 22:41 |
| guitarzan | it's like bizarro day :) | 22:41 |
| hemna | pirate bizarro | 22:41 |
| *** alonma has quit IRC | 22:43 | |
| *** alonma has joined #openstack-cinder | 22:45 | |
| *** david-ly_ is now known as david-lyle | 22:45 | |
| *** jgregor has quit IRC | 22:47 | |
| *** jgregor has joined #openstack-cinder | 22:47 | |
| *** jgregor has quit IRC | 22:48 | |
| *** alonma has quit IRC | 22:49 | |
| *** merooney has joined #openstack-cinder | 22:49 | |
| *** diablo_rojo has quit IRC | 22:50 | |
| *** mriedem is now known as mriedem_away | 22:51 | |
| openstackgerrit | Mitsuhiro Tanino proposed openstack/cinder: Add validation for volume_type of volume object https://review.openstack.org/254414 | 22:51 |
| *** jungleboyj has quit IRC | 22:51 | |
| *** harlowja has quit IRC | 22:51 | |
| *** alonma has joined #openstack-cinder | 22:51 | |
| hemna | w00t | 22:51 |
| hemna | works! | 22:51 |
| hemna | phew | 22:51 |
| *** harlowja has joined #openstack-cinder | 22:51 | |
| hemna | http://paste.openstack.org/show/481401/ | 22:54 |
| scottda | sweet | 22:55 |
| hemna | it only calls remove_export if there are no more attachments on that host. | 22:55 |
| hemna | terminate_connection gets called either way | 22:55 |
| *** alonma has quit IRC | 22:55 | |
| scottda | Brilliant | 22:56 |
| *** esker has joined #openstack-cinder | 22:56 | |
| hemna | I guess I should make a formal patch out of this and push it up for revew | 22:57 |
| hemna | review | 22:57 |
| scottda | And a bug I reckon | 22:57 |
| *** esker has quit IRC | 22:57 | |
| hemna | scottda, https://bugs.launchpad.net/cinder/+bug/1471278 | 22:58 |
| openstack | Launchpad bug 1471278 in Cinder "target removal on detaching volume with multi-attach flag" [Undecided,New] - Assigned to Walt Boring (walter-boring) | 22:58 |
| hemna | that's what started all of this | 22:58 |
| scottda | OH, yeah, that's an oldy but a goody | 22:58 |
| *** esker has joined #openstack-cinder | 22:58 | |
| *** zhipeng has quit IRC | 22:58 | |
| scottda | hemna: Have you been reviewing Illdikov's patches? They seem to keep failing build, so I haven't delved into them.... | 22:59 |
| hemna | and the best part is that the lvm driver isn't changed | 22:59 |
| hemna | phew | 22:59 |
| hemna | scottda, tempest has been giving her fits | 22:59 |
| *** boris-42_ has joined #openstack-cinder | 22:59 | |
| scottda | and none of the drivers need to change, yes? | 22:59 |
| hemna | scottda, correct | 22:59 |
| scottda | so that is very cool. | 22:59 |
| hemna | this even works with the 3PARism | 22:59 |
| *** akshai_ has quit IRC | 22:59 | |
| hemna | since it seems to be the bastard child of the lot | 23:00 |
| scottda | OK, I'll start reviewing those Nova patches. We for sure need the host passed in. But perhaps we can get the connector info as well. | 23:00 |
| *** haomaiwang has quit IRC | 23:01 | |
| *** haomaiwang has joined #openstack-cinder | 23:01 | |
| *** akshai has joined #openstack-cinder | 23:01 | |
| hemna | I suppose it wouldn't be ideal to put the connector change in with the multiattach effort | 23:01 |
| hemna | but sure would be nice to add | 23:01 |
| hemna | https://review.openstack.org/#/c/193134/21/nova/volume/cinder.py | 23:01 |
| hemna | no changes to attach in there | 23:02 |
| *** vgridnev has joined #openstack-cinder | 23:03 | |
| scottda | bummer | 23:04 |
| scottda | But this is only for force-detach, so not critical | 23:04 |
| scottda | Just a dream... | 23:04 |
| *** dave-mccowan has joined #openstack-cinder | 23:07 | |
| hemna | I think that could be next | 23:08 |
| hemna | just need a new column in the attachment table to track the connector per attachment. | 23:08 |
| *** dustins has quit IRC | 23:09 | |
| *** martyturner has quit IRC | 23:13 | |
| *** dustins has joined #openstack-cinder | 23:14 | |
| *** alonma has joined #openstack-cinder | 23:19 | |
| *** vgridnev has quit IRC | 23:19 | |
| *** xyang has quit IRC | 23:22 | |
| *** alonma has quit IRC | 23:24 | |
| *** xyang1 has quit IRC | 23:25 | |
| *** briancurtin has joined #openstack-cinder | 23:29 | |
| *** dilip has joined #openstack-cinder | 23:32 | |
| *** haomaiwang has quit IRC | 23:32 | |
| *** haomaiwang has joined #openstack-cinder | 23:32 | |
| *** ntpttr has left #openstack-cinder | 23:34 | |
| *** ianbrown has joined #openstack-cinder | 23:36 | |
| *** ianbrown has quit IRC | 23:36 | |
| *** haomaiwang has quit IRC | 23:37 | |
| *** rcernin has joined #openstack-cinder | 23:38 | |
| *** takedakn has joined #openstack-cinder | 23:41 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder: EMC VMAX - get_short_host_name not called in find_device_number https://review.openstack.org/252093 | 23:42 |
| *** sgotliv_ has quit IRC | 23:48 | |
| *** patrickeast has joined #openstack-cinder | 23:51 | |
| *** ianbrown has joined #openstack-cinder | 23:55 | |
| *** [1]Thelo has joined #openstack-cinder | 23:57 | |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: WIP don't call remove_export with attachments left https://review.openstack.org/255595 | 23:58 |
| *** Thelo has quit IRC | 23:59 | |
| *** [1]Thelo is now known as Thelo | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!