| *** primechuck has quit IRC | 00:01 | |
| *** tsekiyama has quit IRC | 00:01 | |
| *** DCWillia_ has quit IRC | 00:02 | |
| *** annegentle has joined #openstack-cinder | 00:03 | |
| *** swamireddy1 has joined #openstack-cinder | 00:05 | |
| *** annegentle has quit IRC | 00:08 | |
| *** nikeshm has joined #openstack-cinder | 00:16 | |
| openstackgerrit | Edwin Wang proposed openstack/cinder: Add iSCSI protocol support for IBM FlashSystem https://review.openstack.org/156206 | 00:21 |
|---|---|---|
| *** MentalRay has joined #openstack-cinder | 00:25 | |
| *** heyun has joined #openstack-cinder | 00:26 | |
| *** setmason has joined #openstack-cinder | 00:27 | |
| *** zhenguo has joined #openstack-cinder | 00:28 | |
| *** IlyaG has quit IRC | 00:28 | |
| *** adurbin_ has quit IRC | 00:31 | |
| *** annegentle has joined #openstack-cinder | 00:39 | |
| *** DCWilliams_VA has joined #openstack-cinder | 00:40 | |
| *** setmason has quit IRC | 00:40 | |
| *** nikeshm has quit IRC | 00:44 | |
| *** IlyaG has joined #openstack-cinder | 00:45 | |
| *** _cjones_ has quit IRC | 00:47 | |
| *** IanGovett1 has joined #openstack-cinder | 00:49 | |
| *** IanGovett has quit IRC | 00:49 | |
| *** leeantho has quit IRC | 00:51 | |
| *** IlyaG has quit IRC | 00:52 | |
| *** daneyon_ has quit IRC | 00:53 | |
| *** annegentle has quit IRC | 00:56 | |
| *** shakamunyi has quit IRC | 01:02 | |
| *** jungleboyj has joined #openstack-cinder | 01:08 | |
| *** DCWilliams_VA has quit IRC | 01:12 | |
| *** primechuck has joined #openstack-cinder | 01:24 | |
| *** Lee1092 has joined #openstack-cinder | 01:25 | |
| *** setmason has joined #openstack-cinder | 01:27 | |
| openstackgerrit | Thang Pham proposed openstack/cinder: WIP: Object pinning https://review.openstack.org/180393 | 01:31 |
| *** dannywilson has quit IRC | 01:31 | |
| *** IlyaG has joined #openstack-cinder | 01:31 | |
| openstackgerrit | Thang Pham proposed openstack/cinder: WIP: Object pinning https://review.openstack.org/180393 | 01:33 |
| openstackgerrit | Thang Pham proposed openstack/cinder: WIP: Object pinning https://review.openstack.org/180393 | 01:36 |
| *** vincent_hou has joined #openstack-cinder | 01:36 | |
| *** Yogi1 has joined #openstack-cinder | 01:41 | |
| *** Apoorva has quit IRC | 01:46 | |
| *** Yogi1 has quit IRC | 01:46 | |
| *** Yogi1 has joined #openstack-cinder | 01:48 | |
| *** _cjones_ has joined #openstack-cinder | 01:48 | |
| *** swamireddy1 has quit IRC | 01:48 | |
| *** vilobhmm1 has quit IRC | 01:56 | |
| *** annegentle has joined #openstack-cinder | 01:56 | |
| *** patrickeast has quit IRC | 01:59 | |
| *** erlon has quit IRC | 02:01 | |
| *** annegentle has quit IRC | 02:02 | |
| *** shakamunyi has joined #openstack-cinder | 02:10 | |
| *** vincent_hou has quit IRC | 02:12 | |
| *** primechuck has quit IRC | 02:13 | |
| *** mriedem has quit IRC | 02:14 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/178410 | 02:14 |
| openstackgerrit | Lin Yang proposed openstack/python-cinderclient: Fix typo in comment message https://review.openstack.org/179020 | 02:20 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/os-brick: Updated from global requirements https://review.openstack.org/180830 | 02:20 |
| *** _cjones_ has quit IRC | 02:27 | |
| *** vincent_hou_ has joined #openstack-cinder | 02:29 | |
| *** vincent_hou_ is now known as vincent_hou | 02:31 | |
| *** mtanino has quit IRC | 02:32 | |
| jgriffith | vincent_hou: given that the feature was written by IBM I would fully expect it to work with Stwz | 02:35 |
| *** heyun has quit IRC | 02:35 | |
| *** bkopilov has quit IRC | 02:35 | |
| *** Yogi1 has quit IRC | 02:38 | |
| *** r-daneel has quit IRC | 02:49 | |
| *** dims_ has quit IRC | 02:51 | |
| *** dims has joined #openstack-cinder | 02:52 | |
| *** jungleboyj has quit IRC | 02:56 | |
| *** annegentle has joined #openstack-cinder | 02:57 | |
| *** setmason has quit IRC | 02:58 | |
| *** setmason has joined #openstack-cinder | 03:01 | |
| *** markvoelker has quit IRC | 03:04 | |
| jgriffith | vincent_hou: hmm... https://www.dropbox.com/s/ozrnfbr52dqctny/Screenshot%202015-05-06%2021.05.32.png?dl=0 | 03:05 |
| jgriffith | vincent_hou: so retype w/migrate once seems to have worked fine as the cleanup is on the LVM volume by name | 03:06 |
| jgriffith | vincent_hou: but then try and retype again.... | 03:06 |
| *** heyun has joined #openstack-cinder | 03:06 | |
| jgriffith | vincent_hou: not only does it say it "worked" | 03:06 |
| jgriffith | vincent_hou: but it lied :) | 03:06 |
| jgriffith | vincent_hou: and it gets worse... | 03:07 |
| jgriffith | vincent_hou: delete the volume | 03:07 |
| jgriffith | vincent_hou: the call goes to the DB and deltes it just fine | 03:07 |
| jgriffith | vincent_hou: HOWEVER it sent the command to the lvm backend (wrong backend) | 03:07 |
| jgriffith | vincent_hou: which the "delete non-existent volume" passes (by design) | 03:08 |
| jgriffith | vincent_hou: so user thinks the volume is deleted/gone | 03:08 |
| jgriffith | vincent_hou: but in reality it's stillon the backend | 03:08 |
| jgriffith | vincent_hou: ^^ That's what smcginnis and tbarron were trying to explain in the bug | 03:08 |
| vincent_hou | jgriffith: Hmm, that is a really bad issue. | 03:09 |
| jgriffith | vincent_hou: yes, I'd agree | 03:09 |
| vincent_hou | jgriffith: Thank you for the testing. It helps me to better understand this issue. | 03:10 |
| jgriffith | vincent_hou: no problem :) | 03:11 |
| jgriffith | vincent_hou: let me know if you're going to keep the bug and owrk on it | 03:11 |
| jgriffith | vincent_hou: otherwise I can look at working on it later this week | 03:11 |
| *** dims has quit IRC | 03:11 | |
| vincent_hou | jgriffith: Sure I will keep on it. | 03:12 |
| jgriffith | vincent_hou: cool, let's touch base tomorrow and see how things are going | 03:12 |
| jgriffith | vincent_hou: I'd like to get this fixed/closed and backported *soon* | 03:12 |
| vincent_hou | jgriffith: OK. | 03:13 |
| jgriffith | vincent_hou: and on that note... I'm going to feed some horses and call it a night :) | 03:13 |
| jgriffith | vincent_hou: take care!! | 03:13 |
| vincent_hou | jgriffith: Thanks. Good night. | 03:14 |
| jgriffith | vincent_hou: although I may be back... this whole WARNING keyston auth.identity thing is REALLY pissing me off at this point | 03:14 |
| vincent_hou | jgriffith: what is the bug number? | 03:15 |
| jgriffith | vincent_hou: for the keystone thing? | 03:15 |
| vincent_hou | jgriffith: Yeah. You said the issue bothers you. | 03:16 |
| vincent_hou | Is it really bad for the release? | 03:16 |
| jgriffith | vincent_hou: http://goo.gl/t0e1dO | 03:16 |
| jgriffith | vincent_hou: thingee and morganfainberg have been working it, think they're just going to remove the warning message in keystoneclient last I heard | 03:17 |
| vincent_hou | jgriffith: OK. | 03:18 |
| morganfainberg | jamielennox: ^^ was that the resulting plan? | 03:18 |
| jamielennox | the fix was merged for that, has there been a cinderclient release lately? | 03:18 |
| jgriffith | jamielennox: haha... nope | 03:18 |
| morganfainberg | Did we release the ksc? | 03:18 |
| jgriffith | jamielennox: but I'd PAY good money to be allowed to do one now :) | 03:18 |
| jamielennox | the last one isn't that old: 2015-04-21 | 03:19 |
| jgriffith | morganfainberg: jamielennox yeah, so I thought we just needed it in KSC? | 03:19 |
| *** setmason has quit IRC | 03:19 | |
| jgriffith | jamielennox: morganfainberg so maybe I was mistaken that I thought this was in progress? | 03:20 |
| jamielennox | https://review.openstack.org/#/c/178439/ | 03:20 |
| morganfainberg | Ahh so just need cinder client. | 03:20 |
| morganfainberg | Cool. | 03:20 |
| jamielennox | so there was an issue on the receiving end, for every one of those requests there was a corresponding failure warning in auth_token middleware which we silenced and released | 03:20 |
| jgriffith | jamielennox: oh, so that's better then, I thought this was only "1/2" of the fix | 03:21 |
| morganfainberg | Yeah ksm is fixed. | 03:21 |
| * morganfainberg should verify a stable ksm was released. | 03:21 | |
| jamielennox | afaik if you check out master then this issue is resolved | 03:21 |
| morganfainberg | Been a busy month... Can't remember if I did stable ksm yet. | 03:21 |
| jamielennox | master cinderclient | 03:21 |
| jgriffith | jamielennox: morganfainberg OK.. SWEET | 03:22 |
| jgriffith | jamielennox: morganfainberg just installed master and indeed, we're good :) | 03:22 |
| jgriffith | jamielennox: morganfainberg sorry for bugging/invoking you on that :) | 03:22 |
| jamielennox | no worries | 03:22 |
| morganfainberg | jgriffith: all good. | 03:22 |
| jgriffith | jamielennox: morganfainberg I'll beg thingee to do a tag/push tomorrow | 03:22 |
| jgriffith | no reason not to now that we've released | 03:23 |
| jgriffith | jamielennox: morganfainberg and on that note, both of you have a fantastic evening | 03:23 |
| * jgriffith has left the building | 03:23 | |
| *** annegentle has quit IRC | 03:25 | |
| thingee | jgriffith, jamielennox: I was waiting on some stuff to land. Looks like it merged today, so I can proceed. | 03:26 |
| morganfainberg | Hehe | 03:26 |
| jgriffith | thingee: SWEEET!!!! | 03:26 |
| jgriffith | what a great day :) | 03:27 |
| *** setmason has joined #openstack-cinder | 03:31 | |
| *** sgotliv has joined #openstack-cinder | 03:31 | |
| *** nikeshm has joined #openstack-cinder | 03:36 | |
| *** IlyaG has quit IRC | 03:37 | |
| *** setmason has quit IRC | 03:38 | |
| *** IanGovett1 has quit IRC | 03:40 | |
| *** lpetrut has joined #openstack-cinder | 03:41 | |
| *** rhefner has quit IRC | 03:44 | |
| *** daneyon has joined #openstack-cinder | 03:45 | |
| *** links has joined #openstack-cinder | 03:45 | |
| *** nikeshm has quit IRC | 03:46 | |
| *** _cjones_ has joined #openstack-cinder | 03:53 | |
| *** MentalRay has quit IRC | 03:53 | |
| *** MentalRay has joined #openstack-cinder | 03:53 | |
| *** ebalduf has joined #openstack-cinder | 03:56 | |
| *** _cjones_ has quit IRC | 03:57 | |
| *** setmason has joined #openstack-cinder | 03:58 | |
| *** shyama has joined #openstack-cinder | 04:00 | |
| *** bkopilov has joined #openstack-cinder | 04:01 | |
| *** ebalduf has quit IRC | 04:01 | |
| *** emagana has quit IRC | 04:03 | |
| *** markvoelker has joined #openstack-cinder | 04:04 | |
| openstackgerrit | wanghao proposed openstack/cinder: Notification with volume and snaphost metadata https://review.openstack.org/180400 | 04:07 |
| *** markvoelker has quit IRC | 04:09 | |
| *** setmason_ has joined #openstack-cinder | 04:12 | |
| *** setmason has quit IRC | 04:13 | |
| *** setmason_ is now known as setmason | 04:13 | |
| openstackgerrit | Mike Perez proposed openstack/python-cinderclient: Update release notes for 1.2.0 and 1.2.1 https://review.openstack.org/180852 | 04:17 |
| thingee | jgriffith: ^ | 04:18 |
| openstackgerrit | Mike Perez proposed openstack/python-cinderclient: Update release notes for 1.2.0 and 1.2.1 https://review.openstack.org/180852 | 04:18 |
| *** swamireddy1 has joined #openstack-cinder | 04:21 | |
| *** shyama has quit IRC | 04:36 | |
| *** lpetrut has quit IRC | 04:40 | |
| *** swamireddy1 has quit IRC | 04:40 | |
| *** sgotliv has quit IRC | 04:40 | |
| *** sgotliv has joined #openstack-cinder | 04:40 | |
| *** swamireddy1 has joined #openstack-cinder | 04:40 | |
| *** emagana has joined #openstack-cinder | 04:41 | |
| *** avishay has joined #openstack-cinder | 04:45 | |
| *** emagana has quit IRC | 04:46 | |
| *** MentalRay has quit IRC | 04:47 | |
| *** shyama has joined #openstack-cinder | 04:47 | |
| *** setmason has quit IRC | 04:53 | |
| *** setmason has joined #openstack-cinder | 04:54 | |
| *** BharatK has quit IRC | 04:56 | |
| *** setmason has quit IRC | 05:00 | |
| *** swamireddy1 has quit IRC | 05:03 | |
| *** coolsvap|afk is now known as coolsvap | 05:03 | |
| *** nkrinner has joined #openstack-cinder | 05:04 | |
| *** markvoelker has joined #openstack-cinder | 05:05 | |
| *** markvoelker has quit IRC | 05:10 | |
| *** BharatK has joined #openstack-cinder | 05:12 | |
| *** swamireddy1 has joined #openstack-cinder | 05:12 | |
| *** vilobhmm1 has joined #openstack-cinder | 05:12 | |
| *** IlyaG has joined #openstack-cinder | 05:21 | |
| *** mdbooth has quit IRC | 05:27 | |
| *** mdbooth has joined #openstack-cinder | 05:31 | |
| *** Longgeek has joined #openstack-cinder | 05:33 | |
| *** emagana has joined #openstack-cinder | 05:35 | |
| *** sgotliv has quit IRC | 05:39 | |
| *** emagana has quit IRC | 05:40 | |
| *** _cjones_ has joined #openstack-cinder | 05:42 | |
| *** tshefi has joined #openstack-cinder | 05:44 | |
| *** _cjones_ has quit IRC | 05:46 | |
| *** markvoelker has joined #openstack-cinder | 05:51 | |
| *** vincent_hou_ has joined #openstack-cinder | 05:53 | |
| *** vincent_hou has quit IRC | 05:54 | |
| *** vincent_hou_ is now known as vincent_hou | 05:54 | |
| *** markvoelker has quit IRC | 05:56 | |
| *** shyama has quit IRC | 05:56 | |
| *** shyama has joined #openstack-cinder | 05:56 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 05:57 |
| *** agarciam has joined #openstack-cinder | 05:58 | |
| *** vilobhmm1 has quit IRC | 05:59 | |
| *** nlevinki has joined #openstack-cinder | 06:02 | |
| *** deepakcs has joined #openstack-cinder | 06:06 | |
| *** Longgeek has quit IRC | 06:06 | |
| *** anshul has joined #openstack-cinder | 06:17 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 06:23 |
| openstackgerrit | wanghao proposed openstack/cinder-specs: Adds support to force-delete backups https://review.openstack.org/98325 | 06:23 |
| *** IlyaG has quit IRC | 06:26 | |
| *** IlyaG has joined #openstack-cinder | 06:26 | |
| *** IlyaG has quit IRC | 06:31 | |
| *** vincent_hou has quit IRC | 06:35 | |
| *** emagana has joined #openstack-cinder | 06:37 | |
| *** emagana has quit IRC | 06:42 | |
| *** alecv has joined #openstack-cinder | 06:45 | |
| *** lpetrut has joined #openstack-cinder | 06:47 | |
| *** markvoelker has joined #openstack-cinder | 06:52 | |
| *** nlevinki has quit IRC | 06:52 | |
| *** shyama has quit IRC | 06:55 | |
| *** dannywilson has joined #openstack-cinder | 06:55 | |
| *** dannywilson has quit IRC | 06:56 | |
| *** markvoelker has quit IRC | 06:56 | |
| *** dannywilson has joined #openstack-cinder | 06:57 | |
| *** ronis has joined #openstack-cinder | 06:58 | |
| *** avishay_ has joined #openstack-cinder | 07:00 | |
| *** avishay has quit IRC | 07:01 | |
| *** ankit_ag has joined #openstack-cinder | 07:01 | |
| *** Longgeek has joined #openstack-cinder | 07:07 | |
| *** alecv has quit IRC | 07:11 | |
| *** alecv has joined #openstack-cinder | 07:13 | |
| *** Longgeek has quit IRC | 07:21 | |
| *** Longgeek has joined #openstack-cinder | 07:21 | |
| *** dulek has joined #openstack-cinder | 07:24 | |
| openstackgerrit | wanghao proposed openstack/cinder: Notification with volume and snaphost metadata https://review.openstack.org/180400 | 07:28 |
| *** avishay_ has quit IRC | 07:29 | |
| *** avishay_ has joined #openstack-cinder | 07:29 | |
| *** _cjones_ has joined #openstack-cinder | 07:30 | |
| *** markus_z has joined #openstack-cinder | 07:31 | |
| *** lpetrut has quit IRC | 07:31 | |
| *** emagana has joined #openstack-cinder | 07:32 | |
| *** _cjones_ has quit IRC | 07:35 | |
| *** emagana has quit IRC | 07:36 | |
| *** takedakn has joined #openstack-cinder | 07:37 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 07:39 |
| *** sgotliv has joined #openstack-cinder | 07:40 | |
| *** jordanP has joined #openstack-cinder | 07:41 | |
| *** lpetrut has joined #openstack-cinder | 07:42 | |
| *** takedakn has quit IRC | 07:44 | |
| *** jistr has joined #openstack-cinder | 07:54 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 08:02 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 08:08 |
| openstackgerrit | Anton Arefiev proposed openstack/python-cinderclient: Add findall server side filtering https://review.openstack.org/180534 | 08:14 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 08:19 |
| *** pck has quit IRC | 08:21 | |
| *** pck has joined #openstack-cinder | 08:21 | |
| *** e0ne has joined #openstack-cinder | 08:23 | |
| *** emagana has joined #openstack-cinder | 08:26 | |
| *** IlyaG has joined #openstack-cinder | 08:27 | |
| *** dannywilson has quit IRC | 08:28 | |
| *** dannywilson has joined #openstack-cinder | 08:30 | |
| *** emagana has quit IRC | 08:31 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 08:51 |
| *** markvoelker has joined #openstack-cinder | 08:53 | |
| openstackgerrit | wanghao proposed openstack/cinder: Notification with volume and snaphost metadata https://review.openstack.org/180400 | 08:58 |
| *** markvoelker has quit IRC | 08:58 | |
| *** e0ne is now known as e0ne_ | 09:03 | |
| *** dannywilson has quit IRC | 09:05 | |
| *** Miouge has quit IRC | 09:08 | |
| *** aix has joined #openstack-cinder | 09:14 | |
| *** _cjones_ has joined #openstack-cinder | 09:19 | |
| *** emagana has joined #openstack-cinder | 09:20 | |
| *** Miouge has joined #openstack-cinder | 09:23 | |
| *** _cjones_ has quit IRC | 09:24 | |
| *** Miouge has quit IRC | 09:24 | |
| openstackgerrit | wanghao proposed openstack/cinder: Fix response when query host detail by host name https://review.openstack.org/162601 | 09:24 |
| *** emagana has quit IRC | 09:25 | |
| openstackgerrit | wanghao proposed openstack/cinder: Fix response when query host detail by host name https://review.openstack.org/162601 | 09:29 |
| wanghao | hi guys: There is a tiny bug need your help to review: https://review.openstack.org/#/c/162601/. Thanks:) | 09:30 |
| *** e0ne_ is now known as e0ne | 09:33 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 09:35 |
| *** dims has joined #openstack-cinder | 09:36 | |
| *** dims has quit IRC | 09:41 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 09:43 |
| openstackgerrit | Tom Barron proposed openstack/cinder: Use nfs_oversub_ratio when reporting pool capacity https://review.openstack.org/180199 | 09:46 |
| *** IanGovett has joined #openstack-cinder | 09:48 | |
| *** markvoelker has joined #openstack-cinder | 09:54 | |
| *** markvoelker has quit IRC | 09:58 | |
| *** dims_ has joined #openstack-cinder | 09:59 | |
| *** dulek has quit IRC | 10:01 | |
| *** dulek has joined #openstack-cinder | 10:02 | |
| *** IanGovett1 has joined #openstack-cinder | 10:02 | |
| *** dulek has quit IRC | 10:03 | |
| *** IanGovett has quit IRC | 10:05 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 10:10 |
| *** emagana has joined #openstack-cinder | 10:14 | |
| *** emagana has quit IRC | 10:18 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 10:25 |
| deepakcs | wanghao, ping | 10:26 |
| *** avishay_ has quit IRC | 10:31 | |
| *** avishay__ has joined #openstack-cinder | 10:32 | |
| *** BharatK has quit IRC | 10:36 | |
| *** swamireddy1 has quit IRC | 10:37 | |
| *** Miouge has joined #openstack-cinder | 10:37 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 10:40 |
| *** heyun has quit IRC | 10:42 | |
| *** coolsvap is now known as coolsvap|afk | 10:43 | |
| *** sgotliv has quit IRC | 10:43 | |
| deepakcs | e0ne, ping | 10:44 |
| e0ne | deepakcs: hi | 10:45 |
| deepakcs | e0ne, Hello, quick q on ur comment @ https://review.openstack.org/#/c/172808/3/cinder/db/sqlalchemy/api.py | 10:45 |
| e0ne | sure | 10:45 |
| *** Miouge has quit IRC | 10:45 | |
| deepakcs | e0ne, should i change to _foo(...).( <next line> filter_by(key=key).( <next line> first()) | 10:46 |
| *** ociuhandu has joined #openstack-cinder | 10:47 | |
| e0ne | deepakcs: i think yes. we try to go away from \ usage. but we have a lot of it for in cinder/db/sqlalchemy/api.py:( | 10:47 |
| e0ne | deepakcs: my -1 was more for 'deleted_at field' update | 10:48 |
| e0ne | i can update my comment | 10:48 |
| deepakcs | e0ne, yeah, i inherited what was before, but i am willing to fix my part, just wanted to check the right syntax | 10:48 |
| *** alexpilotti has joined #openstack-cinder | 10:49 | |
| *** BharatK has joined #openstack-cinder | 10:49 | |
| deepakcs | e0ne, reg: deleted_at field... how come thats not needed for _volume_x_metadata_update ? | 10:49 |
| deepakcs | e0ne, I am new to db changes, so I just followed what volume_admin_metadata api's did, as I wanted to create a similar table | 10:50 |
| *** agarciam has quit IRC | 10:50 | |
| deepakcs | e0ne, and turns out that volume_admin_md doesn't update deleted_at ? | 10:50 |
| *** sgotliv has joined #openstack-cinder | 10:50 | |
| *** alecv has quit IRC | 10:50 | |
| *** agarciam has joined #openstack-cinder | 10:51 | |
| *** cdelatte has quit IRC | 10:51 | |
| e0ne | deepakcs: it's confusing | 10:51 |
| deepakcs | e0ne, what is confusing ? my q or api.py ? | 10:52 |
| e0ne | deepakcs: in general, we don't update deleted_at manually because we use delete() method of the model | 10:52 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 10:52 |
| deepakcs | e0ne, so IIUC , in _update we are updating deleted, hence need to update deleted_at also ? | 10:53 |
| e0ne | deepakcs: but if you somewhy don't use delete() methon you need to do it manually. why not just call model_ref.delete()? | 10:54 |
| *** markvoelker has joined #openstack-cinder | 10:55 | |
| *** Miouge has joined #openstack-cinder | 10:58 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 10:58 |
| *** Longgeek has quit IRC | 10:58 | |
| openstackgerrit | Sean McGinnis proposed openstack/cinder: Logging not using oslo.i18n guidelines (db) https://review.openstack.org/180939 | 10:58 |
| *** Longgeek has joined #openstack-cinder | 10:58 | |
| *** markvoelker has quit IRC | 11:00 | |
| *** alecv has joined #openstack-cinder | 11:00 | |
| *** Miouge has quit IRC | 11:03 | |
| deepakcs | e0ne, as i said i am new to db code, so I just followed what other (similar looking) table did | 11:03 |
| deepakcs | e0ne, model_ref.delete() deletes the whole table or the row associated with model_ref ? | 11:05 |
| *** _cjones_ has joined #openstack-cinder | 11:08 | |
| *** BharatK has quit IRC | 11:08 | |
| *** emagana has joined #openstack-cinder | 11:08 | |
| *** ebalduf has joined #openstack-cinder | 11:10 | |
| *** swamireddy1 has joined #openstack-cinder | 11:10 | |
| *** ankit_ag has quit IRC | 11:11 | |
| *** ankit_ag has joined #openstack-cinder | 11:11 | |
| *** emagana has quit IRC | 11:13 | |
| *** _cjones_ has quit IRC | 11:13 | |
| *** ebalduf has quit IRC | 11:14 | |
| e0ne | deepakcs: actually, it marks row as deleted: https://github.com/openstack/cinder/blob/master/cinder/db/sqlalchemy/models.py#L47 | 11:14 |
| deepakcs | e0ne, replacing \ with () gives syntax error with flake8, i think for now i will leave the \ as is | 11:14 |
| *** swamireddy1 has quit IRC | 11:15 | |
| e0ne | deepakcs: ok, fair enough to leave it as is and make these changes in all file in a separate patch | 11:15 |
| deepakcs | e0ne, :) once i figure what to replace \ with | 11:16 |
| *** avishay__ has quit IRC | 11:18 | |
| *** avishay__ has joined #openstack-cinder | 11:18 | |
| *** BharatK has joined #openstack-cinder | 11:20 | |
| openstackgerrit | Abhijeet Malawade proposed openstack/cinder: Cinder os-attach api returns 500 https://review.openstack.org/178565 | 11:21 |
| openstackgerrit | Abhijeet Malawade proposed openstack/cinder: cinder os-detach api returns 500 https://review.openstack.org/178578 | 11:23 |
| openstackgerrit | Ankit Agrawal proposed openstack/cinder: Fix missing translations for log messages https://review.openstack.org/164692 | 11:25 |
| *** agarciam has quit IRC | 11:25 | |
| *** agarciam has joined #openstack-cinder | 11:25 | |
| *** timcl has joined #openstack-cinder | 11:26 | |
| *** e0ne is now known as e0ne_ | 11:27 | |
| *** links has quit IRC | 11:31 | |
| *** links has joined #openstack-cinder | 11:32 | |
| *** dulek has joined #openstack-cinder | 11:38 | |
| *** marcusvrn has joined #openstack-cinder | 11:41 | |
| *** marcusvrn2 has quit IRC | 11:43 | |
| *** agarciam has quit IRC | 11:49 | |
| *** agarciam has joined #openstack-cinder | 11:49 | |
| *** nlevinki has joined #openstack-cinder | 11:49 | |
| *** avishay__ has quit IRC | 11:50 | |
| *** avishay has joined #openstack-cinder | 11:51 | |
| *** e0ne_ is now known as e0ne | 11:51 | |
| *** openstackgerrit has quit IRC | 11:52 | |
| *** openstackgerrit has joined #openstack-cinder | 11:53 | |
| *** markvoelker has joined #openstack-cinder | 11:56 | |
| *** swamireddy1 has joined #openstack-cinder | 11:57 | |
| openstackgerrit | wanghao proposed openstack/cinder: Notification with volume and snaphost metadata https://review.openstack.org/180400 | 11:58 |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 11:59 |
| *** markvoelker has quit IRC | 12:00 | |
| *** lpabon has joined #openstack-cinder | 12:01 | |
| *** emagana has joined #openstack-cinder | 12:02 | |
| *** emagana has quit IRC | 12:07 | |
| openstackgerrit | yogeshprasad proposed openstack/cinder: Fix the KeyError in CloudByte iSCSI cinder driver https://review.openstack.org/180962 | 12:10 |
| *** ankit_ag has quit IRC | 12:11 | |
| *** akerr has joined #openstack-cinder | 12:12 | |
| *** ankit_ag has joined #openstack-cinder | 12:12 | |
| *** waterkinfe has joined #openstack-cinder | 12:13 | |
| *** anuragpalsule has joined #openstack-cinder | 12:15 | |
| *** markvoelker has joined #openstack-cinder | 12:15 | |
| *** waterkinfe has quit IRC | 12:15 | |
| *** sgotliv has quit IRC | 12:20 | |
| *** jungleboyj has joined #openstack-cinder | 12:23 | |
| *** fthiagogv has joined #openstack-cinder | 12:25 | |
| *** cdelatte has joined #openstack-cinder | 12:28 | |
| openstackgerrit | wanghao proposed openstack/cinder-specs: Adds support to force-delete backups https://review.openstack.org/98325 | 12:35 |
| *** Yogi1 has joined #openstack-cinder | 12:36 | |
| *** bkopilov has quit IRC | 12:39 | |
| *** marcusvrn1 has joined #openstack-cinder | 12:42 | |
| *** fthiagogv has quit IRC | 12:43 | |
| *** swamireddy2 has joined #openstack-cinder | 12:44 | |
| *** swamireddy1 has quit IRC | 12:44 | |
| *** marcusvrn has quit IRC | 12:45 | |
| *** dencaval has joined #openstack-cinder | 12:45 | |
| *** fthiagogv has joined #openstack-cinder | 12:46 | |
| *** sgotliv has joined #openstack-cinder | 12:49 | |
| *** dulek has quit IRC | 12:50 | |
| *** dims_ has quit IRC | 12:51 | |
| *** dims has joined #openstack-cinder | 12:51 | |
| *** bswartz has quit IRC | 12:53 | |
| *** dulek has joined #openstack-cinder | 12:54 | |
| *** deepakcs has quit IRC | 12:54 | |
| *** emagana has joined #openstack-cinder | 12:57 | |
| *** _cjones_ has joined #openstack-cinder | 12:57 | |
| *** IanGovett1 has quit IRC | 12:58 | |
| *** crose has joined #openstack-cinder | 12:59 | |
| *** crose has quit IRC | 12:59 | |
| *** IanGovett has joined #openstack-cinder | 13:00 | |
| *** emagana has quit IRC | 13:01 | |
| *** merooney has joined #openstack-cinder | 13:01 | |
| *** mriedem has joined #openstack-cinder | 13:02 | |
| *** _cjones_ has quit IRC | 13:02 | |
| *** ctina__ has joined #openstack-cinder | 13:02 | |
| *** IanGovett has quit IRC | 13:10 | |
| openstackgerrit | Deepak C Shetty proposed openstack/cinder: WIP: Cinder part of volume snapshot improvements https://review.openstack.org/172808 | 13:12 |
| *** IanGovett has joined #openstack-cinder | 13:13 | |
| openstackgerrit | Gorka Eguileor proposed openstack/cinder: Follow i18n guidelines in LIO target https://review.openstack.org/180520 | 13:14 |
| *** pwehrle has joined #openstack-cinder | 13:18 | |
| *** david-lyle has quit IRC | 13:19 | |
| *** merooney has quit IRC | 13:20 | |
| *** nlevinki has quit IRC | 13:21 | |
| *** openstackstatus has quit IRC | 13:25 | |
| *** openstackstatus has joined #openstack-cinder | 13:25 | |
| *** ChanServ sets mode: +v openstackstatus | 13:25 | |
| openstackgerrit | Ph. Marek proposed openstack/cinder: Re-add DRBD driver. https://review.openstack.org/178573 | 13:26 |
| *** primechuck has joined #openstack-cinder | 13:27 | |
| *** eharney has quit IRC | 13:28 | |
| *** dustins has joined #openstack-cinder | 13:28 | |
| *** bswartz has joined #openstack-cinder | 13:29 | |
| *** changbl has quit IRC | 13:33 | |
| *** jungleboyj has quit IRC | 13:33 | |
| *** e0ne is now known as e0ne_ | 13:34 | |
| *** rushil has joined #openstack-cinder | 13:35 | |
| *** nlevinki has joined #openstack-cinder | 13:36 | |
| *** e0ne_ is now known as e0ne | 13:37 | |
| *** BharatK has quit IRC | 13:38 | |
| *** chlong has joined #openstack-cinder | 13:39 | |
| *** marcusvrn1 has quit IRC | 13:41 | |
| *** marcusvrn has joined #openstack-cinder | 13:44 | |
| *** kashyap has joined #openstack-cinder | 13:44 | |
| *** nlevinki has quit IRC | 13:47 | |
| *** nlevinki has joined #openstack-cinder | 13:47 | |
| *** rushil has quit IRC | 13:48 | |
| *** r-daneel has joined #openstack-cinder | 13:49 | |
| *** esker has joined #openstack-cinder | 13:50 | |
| *** ebalduf has joined #openstack-cinder | 13:50 | |
| kashyap | Hi folks, an admin question: is it possible to migrate a Cinder volume from one LVM backend to another LVM backend on the _same_ host? | 13:50 |
| *** jamielennox is now known as jamielennox|away | 13:51 | |
| *** setmason has joined #openstack-cinder | 13:52 | |
| *** swamireddy2 has quit IRC | 13:52 | |
| *** fanyaohong has joined #openstack-cinder | 13:54 | |
| jbernard | kashyap: yes.. | 13:54 |
| jbernard | kashyap: at least once :) | 13:54 |
| *** links has quit IRC | 13:54 | |
| *** IanGovett1 has joined #openstack-cinder | 13:55 | |
| kashyap | jbernard: Ah, thanks for confirming, you saved time from digging into code (although that'd be a more useful exercise) :-) | 13:55 |
| kashyap | jbernard: I've setup a local DevStack instance w/ two LVM backends | 13:55 |
| jbernard | kashyap: here's a command that should get you close: 'cinder retype --migration-policy on-demand <volume> <backend>' | 13:55 |
| smcginnis | kashyap: There is a patch out there right now to address an issue with the aftermath of a migration though: https://review.openstack.org/#/c/180873/ | 13:56 |
| smcginnis | kashyap: As it is right now, there may be some manual cleanup to delete the volume after migration. | 13:56 |
| smcginnis | See the referenced bug for more details. | 13:56 |
| jbernard | smcginnis: thanks | 13:57 |
| smcginnis | jbernard: np! | 13:57 |
| *** IanGovett has quit IRC | 13:57 | |
| kashyap | scottda: jbernard: Thanks, both for the details and URLs. I don't normally test Cinder in my DevStack workflow, now I've made it part of it :-). | 13:57 |
| kashyap | smcginnis: Is there any place where I can read about this aftermath cleanup? | 13:58 |
| smcginnis | kashyap: Just the comments in the bug. | 13:58 |
| *** geguileo has quit IRC | 13:59 | |
| smcginnis | kashyap: The main issue being after it's migrated the driver can't find it anymore. | 13:59 |
| *** delattec has joined #openstack-cinder | 13:59 | |
| smcginnis | kashyap: So when you try to delete the volume the driver can't find it and thinks it's already gone. | 13:59 |
| smcginnis | kashyap: So it returns success. | 13:59 |
| smcginnis | kashyap: But the volume is still out there, so you would need to manually run the LVM commands to clean it up. | 14:00 |
| smcginnis | kashyap: Also applies to any other operations beside delete where the driver needs to modify the volume. | 14:00 |
| kashyap | smcginnis: Ah, I see, thanks for this info. I created the 2nd LVM backend, but didn't create the VG, so Cinder complained after a restart. So I just did the sequence of: | 14:01 |
| kashyap | $ truncate -s 10250M /home/kashyapc/src/cloud/data/stack-volumes-lvmdriver-2-backing-file | 14:01 |
| kashyap | $ sudo losetup -f --show /home/kashyapc/src/cloud/data/stack-volumes-lvmdriver-2-backing-file | 14:01 |
| kashyap | $ sudo vgcreate stack-volumes-lvmdriver-2 /dev/loop2 | 14:01 |
| kashyap | Following that, I did this - https://kashyapc.fedorapeople.org/virt/openstack/cinder-lvm-backends-test.txt | 14:01 |
| kashyap | So, after restartin Cinder services (presumably just 'c-vol' would suffice I guess), `cinder-manage host list` should enumerate both the backends, I guess? | 14:02 |
| *** cdelatte has quit IRC | 14:02 | |
| *** thangp has joined #openstack-cinder | 14:03 | |
| *** nkrinner has quit IRC | 14:04 | |
| kashyap | I'll investigate the details, thanks all for the responses. | 14:04 |
| smcginnis | kashyap: Looks right. Though to be honest, other than knowing about that bug I haven't done much with the LVM driver myself. | 14:05 |
| *** geguileo has joined #openstack-cinder | 14:05 | |
| kashyap | smcginnis: No problem. Me neither, I mostly work with qcow2 files and backing chains. | 14:05 |
| *** eharney has joined #openstack-cinder | 14:06 | |
| *** anuragpalsule has quit IRC | 14:07 | |
| *** IlyaG has quit IRC | 14:07 | |
| *** IlyaG has joined #openstack-cinder | 14:08 | |
| *** IlyaG has quit IRC | 14:12 | |
| *** IlyaG has joined #openstack-cinder | 14:12 | |
| *** atan8 has joined #openstack-cinder | 14:14 | |
| *** Yogi1 has quit IRC | 14:14 | |
| *** ankit_ag has quit IRC | 14:14 | |
| *** rushil has joined #openstack-cinder | 14:16 | |
| *** setmason has quit IRC | 14:16 | |
| *** agarciam has quit IRC | 14:18 | |
| openstackgerrit | Ph. Marek proposed openstack/cinder: Re-add DRBD driver. https://review.openstack.org/178573 | 14:20 |
| *** hodos has joined #openstack-cinder | 14:20 | |
| *** bkopilov has joined #openstack-cinder | 14:20 | |
| *** avishay_ has joined #openstack-cinder | 14:27 | |
| *** avishay has quit IRC | 14:27 | |
| openstackgerrit | Eric Harney proposed openstack/cinder: LIO: Use rtslib property instead of private method https://review.openstack.org/165142 | 14:28 |
| openstackgerrit | Eric Harney proposed openstack/cinder: LIO: Enable iSER for IPv6 https://review.openstack.org/165143 | 14:28 |
| *** geguileo has quit IRC | 14:30 | |
| *** geguileo has joined #openstack-cinder | 14:34 | |
| *** jungleboyj has joined #openstack-cinder | 14:34 | |
| *** setmason has joined #openstack-cinder | 14:39 | |
| *** emagana has joined #openstack-cinder | 14:41 | |
| *** merooney has joined #openstack-cinder | 14:42 | |
| *** BharatK has joined #openstack-cinder | 14:44 | |
| thangp | DuncanT: hi, do you have time to go over what you plan for rpc version compatibility? | 14:45 |
| *** _cjones_ has joined #openstack-cinder | 14:45 | |
| jordanP | someone with a glusterfs at hand here ? | 14:46 |
| jordanP | when I "cinder create --snapshot-id $snap_id $volume_size", the volume size seems not honored | 14:46 |
| jordanP | although cinder list reports the correct size, on the backend the file is not correctly sized (I am using raw image) | 14:47 |
| hodos | why tox -e all -- 'tempest.api.volume.(?!.*(test_volume_upload|test_volume_backup))' runs only 243 tests ? | 14:47 |
| *** crose has joined #openstack-cinder | 14:48 | |
| openstackgerrit | yogeshprasad proposed openstack/cinder: Fix the KeyError in CloudByte iSCSI cinder driver https://review.openstack.org/180962 | 14:48 |
| *** MentalRay has joined #openstack-cinder | 14:49 | |
| *** ronis has quit IRC | 14:49 | |
| *** _cjones_ has quit IRC | 14:50 | |
| *** hemnafk is now known as hemna | 14:52 | |
| jungleboyj | hemna: Morning sir. | 14:53 |
| hemna | mornin | 14:53 |
| hemna | it's raining here in CA! | 14:53 |
| hemna | :P | 14:53 |
| e0ne | hemna: it's good sunny evening;) | 14:54 |
| hemna | :) | 14:54 |
| jungleboyj | :-) We are waiting to see if severe weather develops this afternoon. :-) | 14:54 |
| jgriffith | hemna: now, if it will do that for about the next 90 days straight, you guys will be all set! | 14:54 |
| jungleboyj | hemna: Oh master of FC do you know if Brocade has any plan to pick the virtual fabrics support back up? | 14:55 |
| hemna | jgriffith, +1! | 14:55 |
| hemna | jungleboyj, afaik, I think they were going to. | 14:55 |
| *** nlevinki has quit IRC | 14:55 | |
| hemna | jungleboyj, they also were going to update their drive to use an http API instead of ssh | 14:55 |
| *** ociuhandu_ has joined #openstack-cinder | 14:56 | |
| *** ociuhandu has quit IRC | 14:56 | |
| *** ociuhandu_ is now known as ociuhandu | 14:56 | |
| *** anshul has quit IRC | 14:57 | |
| jungleboyj | hemna: Ok, cool. Have some people here looking at it. Going to see if we can give Brocade a push on that. | 14:57 |
| hemna | yah that might be good | 14:57 |
| *** crose has quit IRC | 14:58 | |
| *** xyang1 has joined #openstack-cinder | 14:58 | |
| *** juzuluag has joined #openstack-cinder | 15:00 | |
| jgriffith | smcginnis: tbarron did you see my conversation last night WRT migration ? | 15:01 |
| smcginnis | jgriffith: Yea, thanks! | 15:01 |
| jgriffith | cool | 15:01 |
| smcginnis | jgriffith: Looks like he has a patch out there for it. | 15:01 |
| jgriffith | meh | 15:01 |
| tbarron | jgriffith: yes, thanks on that | 15:01 |
| smcginnis | jgriffith: Haven't tested it yet with our driver. Didn't realize there was a method in there already. | 15:01 |
| jgriffith | smcginnis: well, there wasn't.. his patch added one | 15:02 |
| jgriffith | Unless you're looking at something else | 15:03 |
| DuncanT | thangp: I'm out at the moment, can do in about two hours if that helps? | 15:03 |
| jgriffith | DuncanT: "out" where... you seem "here" to me :) | 15:03 |
| thangp | DuncanT: sure, thx | 15:03 |
| smcginnis | jgriffith: https://github.com/openstack/cinder/blob/master/cinder/volume/driver.py#L1345 | 15:03 |
| e0ne | hemna: i'll publish early poc of cinder agent tonight. one 1 call (get connector initiator works) | 15:03 |
| hemna | sweet | 15:03 |
| smcginnis | jgriffith: He just made some updates. | 15:03 |
| hemna | nice job | 15:03 |
| e0ne | hemna: i'm hitting with keystonemiddleware:( | 15:04 |
| jgriffith | smcginnis: ahh.. indeed | 15:04 |
| e0ne | it's very easy to implement new features. but it takes a lot of time to integrate with openstack infrastructure. need to know about logging, configuration, etc | 15:05 |
| *** annegentle has joined #openstack-cinder | 15:09 | |
| *** tshefi has quit IRC | 15:11 | |
| hemna | e0ne, so are we sure we actually need keystone though? | 15:11 |
| hemna | this is most likely targeted as an ironic BM node install, so I'm not sure we want that agent talking to the deployed keystone service ? | 15:12 |
| e0ne | hemna: may be not. i'll share current code with you once it pass pep8 | 15:12 |
| *** pwehrle has quit IRC | 15:13 | |
| hemna | It's something we should talk about I guess | 15:13 |
| *** annegentle has quit IRC | 15:14 | |
| smcginnis | jgriffith: Oh, this is funny. https://github.com/openstack/cinder/blame/master/cinder/volume/driver.py#L1345 | 15:14 |
| smcginnis | hehe | 15:14 |
| *** annegentle has joined #openstack-cinder | 15:14 | |
| e0ne | agree. if we'll deploy it inside ironic instance, it won't be able to access keystone in some cases | 15:15 |
| hodos | jgriffith: hi john. what string do you use when start tempest? tox -e all -- vollume | 15:15 |
| jgriffith | hodos: yeah, but I spell volume with one 'l' :) | 15:15 |
| e0ne | :) | 15:16 |
| hodos | jgriffith: sure. for some reason my test count in this case is 244 | 15:16 |
| hodos | hodos: not 304 | 15:16 |
| jgriffith | hodos: missing services maybe? | 15:16 |
| jgriffith | hodos: what's your local.conf look like? | 15:16 |
| hodos | jgriffith: kinda like yours from sos-ci repo | 15:16 |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: WIP: Remove Brick from cinder codebase https://review.openstack.org/155552 | 15:17 |
| jgriffith | hodos: hmm... do you have the testr_results.html file? | 15:17 |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: Remove Brick from cinder codebase https://review.openstack.org/155552 | 15:17 |
| hodos | jgriffith: i can see that region_list and quotas are omitted | 15:17 |
| hodos | jgriffith: http://140.174.232.106/refs-changes-43-170143-58/testr_results.html | 15:18 |
| hodos | jgriffith: i skipped backup and upload i think. maybe it's the reason ? | 15:18 |
| jgriffith | hodos: hmm | 15:19 |
| *** JayJ_ has joined #openstack-cinder | 15:21 | |
| jgriffith | hodos: strage there are 60 tests not even showing up | 15:21 |
| jgriffith | hodos: lemme double check my tox arg | 15:21 |
| jgriffith | hodos: yeah, just "tox -e all -- volume" | 15:22 |
| *** harlowja_at_home has joined #openstack-cinder | 15:22 | |
| *** garthb has joined #openstack-cinder | 15:22 | |
| *** dannywilson has joined #openstack-cinder | 15:22 | |
| jgriffith | hodos: and you of course have: http://hastebin.com/fidohezige.ini in your local.conf | 15:23 |
| jgriffith | hodos: or your equivalent of it | 15:23 |
| jgriffith | hodos: what I'm not sure I get is your count is different | 15:24 |
| jgriffith | hodos: it's not even a matter of skips or anything | 15:24 |
| jgriffith | hodos: oh.. wait | 15:25 |
| jgriffith | hodos: what's enabled-services have in your local.conf file? | 15:25 |
| smcginnis | hodos: Do you limit ENABLED_SERVICES? | 15:25 |
| hodos | jgriffith: ok, sorry for disturbance | 15:25 |
| jgriffith | hodos: assuming that means a missing service perhaps? | 15:25 |
| hodos | smcginnis: jgriffith: sorry guys; my string looks like this | 15:25 |
| jgriffith | hodos: BTW, you should publish the local.conf file with the results as well | 15:26 |
| hodos | jgriffith: smcginnis: tox -e all -- 'tempest.api.volume.(?!.*(test_volume_upload|test_volume_backup))' and some tests are NOT tempest.api | 15:26 |
| hodos | jgriffith: that's the reason | 15:26 |
| jgriffith | hodos: yep | 15:26 |
| smcginnis | hodos: Yeah, I used to do similar (api.volume) but had to expand that | 15:26 |
| jgriffith | hodos: that's the "old" test | 15:26 |
| *** jdurgin1 has joined #openstack-cinder | 15:31 | |
| *** harlowja_at_home has quit IRC | 15:35 | |
| *** rushil has quit IRC | 15:39 | |
| thingee | jgriffith: https://review.openstack.org/#/c/180852/ | 15:39 |
| *** daneyon_ has joined #openstack-cinder | 15:41 | |
| jgriffith | thingee: Rock ON! | 15:41 |
| *** zhenguo has quit IRC | 15:44 | |
| *** daneyon has quit IRC | 15:44 | |
| *** fthiagogv_ has joined #openstack-cinder | 15:46 | |
| e0ne | hemna: https://github.com/e0ne/cinder-agent-poc | 15:47 |
| hemna | e0ne, coolio | 15:47 |
| hemna | I'll take a look at it later today. | 15:47 |
| hemna | trying to update my nova os-brick patch | 15:47 |
| *** LarsN has joined #openstack-cinder | 15:48 | |
| *** lpetrut has quit IRC | 15:48 | |
| e0ne | thanks. any feedback is welcome | 15:48 |
| LarsN | https://wiki.openstack.org/wiki/Cinder/blueprints/multi-attach-volume | 15:48 |
| e0ne | it's really draft code | 15:48 |
| jgriffith | smcginnis: so that's not really the problem | 15:48 |
| *** fthiagogv has quit IRC | 15:48 | |
| LarsN | is that Blueprint for Multi-Attach on the roadmap? | 15:48 |
| smcginnis | jgriffith: No? | 15:48 |
| jgriffith | smcginnis: the real problem is that there's ZERO scheudler involvement in that migrate/retype op | 15:48 |
| thingee | hemna: link to os-brick nova patch? | 15:48 |
| jgriffith | smcginnis: so the rpc call is sent directly to the backend/driver | 15:48 |
| *** markus_z has quit IRC | 15:49 | |
| smcginnis | jgriffith: Is it missing something then by not going through the scheduler? | 15:49 |
| hemna | thingee, https://review.openstack.org/#/c/175569/ | 15:49 |
| jgriffith | smcginnis: and most of the drivers "assume" that if the request got to them the scheduler did it's job and gave it something it can do | 15:49 |
| tbarron | ping winston-d_zzZ, xyang1: you two may not be on now but I just want to see if we can talk about future nfs overprovisioning and integration with the scheduler overprovisioning work you two have done. | 15:49 |
| jgriffith | the result is they typically find this to be a noop and just return True | 15:49 |
| hemna | thingee, I'm updating it to fix the unit tests and also to migrate to use the other connectors in os-brick (FC, iSER, etc) | 15:49 |
| smcginnis | jgriffith: So there's a gap if the migration fails? I'm still not clear. | 15:49 |
| jgriffith | in the manager the code say "ok" it's done | 15:49 |
| jgriffith | smcginnis: no, the gap is migration is never even called | 15:50 |
| *** fthiagogv_ has quit IRC | 15:50 | |
| smcginnis | jgriffith: Oh!? | 15:50 |
| *** mtanino has joined #openstack-cinder | 15:50 | |
| jgriffith | smcginnis: yeah, because we just blindly send the retype down | 15:50 |
| *** fthiagogv has joined #openstack-cinder | 15:50 | |
| tbarron | nfs overprovisioning pre-dated the scheduler based overprovisioning introduced in kilo | 15:50 |
| jgriffith | smcginnis: no regard for whehter the backend is capable or not | 15:50 |
| thingee | hemna: right on! | 15:51 |
| jgriffith | smcginnis: try this lille test: | 15:51 |
| hemna | I want to get this done in L1 if possible | 15:51 |
| tbarron | and relies on the driver over-reporting free space back up to the scheduler (via manager of course) | 15:51 |
| *** eglynn has joined #openstack-cinder | 15:51 | |
| jgriffith | smcginnis: stub a migrate complete method in your driver and set a break point | 15:51 |
| thingee | e0ne: https://review.openstack.org/#/c/180852/ :) | 15:51 |
| jgriffith | smcginnis: create a volume on your backend | 15:51 |
| jgriffith | smcginnis: then retype it to lvm with migration on-demand | 15:51 |
| thingee | xyang1: https://review.openstack.org/#/c/180852/ | 15:51 |
| xyang1 | tbarron: I am in a call now. Should we discuss at the summit about this? This is for future, right? | 15:51 |
| tbarron | we should either deprecate the nfs overprovisioning options or re-implement them on top of the new scheduler-based overprovisioning. | 15:52 |
| e0ne | thingee: great! thanks a lot for doing it and notifiing me | 15:52 |
| xyang1 | thingee: Nice! | 15:52 |
| tbarron | xyang1: yes, but I think we should get it done in liberty. | 15:52 |
| tbarron | yes future in that sense ^^^^ | 15:52 |
| *** mtanino_ has joined #openstack-cinder | 15:52 | |
| smcginnis | jgriffith: I need to get my stack up again. Sounds like I've some testing to do. | 15:52 |
| xyang1 | tbarron: Summit is for liberty | 15:52 |
| tbarron | xyang1: ok, we'll talk then | 15:53 |
| tbarron | xyang1: thingee: eharney: winston-1: winston-d_zzZ: could be a sprint, or maybe just a coffee break | 15:54 |
| thingee | we really should get the sprint thing organized | 15:54 |
| thingee | We have the full day friday for it | 15:55 |
| tbarron | thingee: :-) | 15:55 |
| thingee | I was imagining the first half would be people talking/getting ideas from discussion from the group. | 15:55 |
| thingee | the second half is actual doing stuff. POC's | 15:55 |
| *** mtanino has quit IRC | 15:55 | |
| eharney | tbarron: i think we've already got a reasonable start on this... with https://review.openstack.org/#/c/149765/ landed, i think the only task is to make sure the driver reports the requisite stats? | 15:55 |
| thingee | is it possible? or will we be fried? | 15:55 |
| tbarron | thingee: that makes sense to me | 15:55 |
| tbarron | eharney: great, I'll take a look. | 15:56 |
| *** vincent_hou has joined #openstack-cinder | 15:56 | |
| xyang1 | tbarron: Sounds good | 15:56 |
| eharney | tbarron: i think we just need something like https://review.openstack.org/#/c/162718/ for the NFS driver, i'm not sure there's a lot else to figure out on that front | 15:56 |
| vincent_hou | jgriffith: Ping again :-) | 15:56 |
| jgriffith | vincent_hou: oh, you're still up :) | 15:57 |
| *** marcusvrn has quit IRC | 15:57 | |
| vincent_hou | jgriffith: Y. I have submitted a patch for the issue we discussed. | 15:57 |
| vincent_hou | https://review.openstack.org/#/c/180873 | 15:58 |
| tbarron | eharney: do you have an opinion as to whether we should deprecate 'nfs_oversub_ratio' and just ues 'max_oversubscription_ratio', or whether | 15:58 |
| *** avishay_ has quit IRC | 15:58 | |
| tbarron | we should support the old option using the new scheduler mechanics? | 15:58 |
| *** daneyon_ has quit IRC | 15:58 | |
| LarsN | Is the ability to attach to a Cinder Volume from multiple instances a feature that exists, or might exist soon? | 15:58 |
| tbarron | I'm kinda inclined towards deprecate, and document that one kind of "thin" provisioning with NFS etc. involves sparse volume creation. | 15:59 |
| eharney | tbarron: yes, if we use the NFS driver configuration method above, there's no reason to not move toward the new option and deprecate the nfs_oversub_ratio one | 15:59 |
| tbarron | eharney: +1 | 15:59 |
| *** daneyon has joined #openstack-cinder | 15:59 | |
| *** mtanino_ has quit IRC | 15:59 | |
| *** fanyaohong has quit IRC | 15:59 | |
| eharney | tbarron: part of my goal with that configuration change was to remove the pseudo-scheduling code that's in the NFS driver anyway, which this lines up with | 15:59 |
| *** marcusvrn has joined #openstack-cinder | 16:00 | |
| *** rmesta has joined #openstack-cinder | 16:00 | |
| tbarron | so as far as the scheduler is concerned, we'll report that the pool has "thin" provisioning if either it has sparse-volumes (default for NFS) | 16:00 |
| tbarron | or if they aren't sparse but there is some vendor-specific thin provisioning going on. | 16:00 |
| tbarron | eharney: I am in complete agreement about eliminating pseudo-scheduling code from the NFS driver. | 16:01 |
| eharney | tbarron: yes i think so | 16:01 |
| *** _cjones_ has joined #openstack-cinder | 16:01 | |
| tbarron | eharney: thanks for the review references, I'll start up work on the generic NFS and NetApp NFS fronts that corresponds - unless you are already doing generic NFS. | 16:02 |
| jgriffith | winston-1: you around? | 16:02 |
| eharney | tbarron: i'm not but i think there is probably a starting point or some overlap in the glusterfs patch referenced above, which puts some of the code in remotefs for reuse | 16:02 |
| *** Yogi1 has joined #openstack-cinder | 16:03 | |
| tbarron | eharney: my plan of action was: #1) steal what I can (#2) iterate on #1, etc. | 16:03 |
| eharney | sounds good :) | 16:03 |
| tbarron | :-) | 16:03 |
| *** bkopilov has quit IRC | 16:04 | |
| *** e0ne is now known as e0ne_ | 16:08 | |
| *** bkopilov has joined #openstack-cinder | 16:09 | |
| vincent_hou | jgriffith: can you help to look at https://review.openstack.org/#/c/180873/? | 16:10 |
| jgriffith | vincent_hou: yes | 16:11 |
| jgriffith | vincent_hou: so that's part of the problem and looks ok so far | 16:12 |
| jgriffith | vincent_hou: but there's another bigger problem | 16:12 |
| jgriffith | vincent_hou: I'm working on right now | 16:12 |
| jgriffith | vincent_hou: we never check to see if the backend is able to support the type when we send the retype command down | 16:12 |
| jgriffith | vincent_hou: there's no scheduling checks at all :( | 16:12 |
| tbarron | jgriffith: +1 | 16:13 |
| tbarron | jgriffith: now I know why you were asking if winston-1 is around. | 16:13 |
| *** _cjones_ has quit IRC | 16:14 | |
| jgriffith | tbarron: yeah, working on a temporary hack to check volume_backend_name in the types | 16:14 |
| jgriffith | tbarron: then, follow up with a full on scheduler/capability check | 16:15 |
| *** lpetrut has joined #openstack-cinder | 16:15 | |
| tbarron | jgriffith: makes sense to me | 16:15 |
| jgriffith | tbarron: this sort of thing has come up before, wanted to make sure there wasn't an easy hook to that call burried somewhere that I could just "use" :) | 16:16 |
| *** vincent_hou_ has joined #openstack-cinder | 16:16 | |
| vincent_hou_ | jgriffith: It seems migration really has a lot of work to do. If you think there needs to do any improvement to it, you can add here https://etherpad.openstack.org/p/volume-migration-improvement. | 16:16 |
| jgriffith | vincent_hou_: oh.. handy | 16:16 |
| *** e0ne_ is now known as e0ne | 16:17 | |
| *** vincent_hou has quit IRC | 16:17 | |
| *** vincent_hou_ is now known as vincent_hou | 16:17 | |
| vincent_hou | jgriffith: I would like to make migration more stable in the coming release. Already put some items in this page. | 16:18 |
| *** rushil has joined #openstack-cinder | 16:18 | |
| xyang1 | tbarron: How are you planning to move forward with your current patch? Get it in first, and then move to new mechanism later in liberty? Or are you going to redo the current patch? | 16:18 |
| tbarron | xyang1: I want to go ahead with the current patch so I can cherry-pick and backport to the customer, who won't see liberty code for a long time. | 16:19 |
| xyang1 | tbarron: Ok, makes sense to me | 16:19 |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: Remove Brick from cinder codebase https://review.openstack.org/155552 | 16:19 |
| tbarron | But then I want to re-do it in liberty so that the driver no longer reports apparent free space up to the scheduler | 16:20 |
| xyang1 | tbarron: Sure | 16:20 |
| tbarron | we can document that nfs_oversub_ratio is no longer supported and how to get the same thing done with the new options. | 16:20 |
| *** aviramb has joined #openstack-cinder | 16:20 | |
| xyang1 | tbarron: Ok | 16:21 |
| tbarron | it looks like eharney has already been doing this for glusterfs and remotefs, so we'll try to leverage that work for consistent approach. | 16:21 |
| xyang1 | tbarron: Right | 16:21 |
| tbarron | so on the current review I think I just have to fix a pylint issue and you and Winston will be OK since you know I'm not trying to have that code live on. | 16:22 |
| *** _cjones_ has joined #openstack-cinder | 16:23 | |
| xyang1 | tbarron: I am ok with your plan | 16:23 |
| tbarron | To make pylint happy I have to do a literal copy of the same method into two libraries and do a NotImplemented version in a third :-) | 16:23 |
| *** changbl has joined #openstack-cinder | 16:23 | |
| tbarron | at run-time the method resolution just worked :-) | 16:23 |
| xyang1 | Have fun:) | 16:23 |
| tbarron | xyang1: will do, and thanks for the help. | 16:24 |
| xyang1 | tbarron: Welcome | 16:24 |
| *** mtanino has joined #openstack-cinder | 16:25 | |
| *** vincent_hou_ has joined #openstack-cinder | 16:25 | |
| *** Apoorva has joined #openstack-cinder | 16:26 | |
| *** vincent_hou has quit IRC | 16:27 | |
| *** vincent_hou_ is now known as vincent_hou | 16:27 | |
| openstackgerrit | Vincent Hou proposed openstack/cinder: Implement the update_migrated_volume for the drivers https://review.openstack.org/180873 | 16:27 |
| *** shakamunyi has quit IRC | 16:29 | |
| mtanino | vincent_hou: hi, | 16:30 |
| vincent_hou | mtanino: Hi | 16:30 |
| mtanino | I'm looking the https://review.openstack.org/#/c/180873/ :) | 16:30 |
| mtanino | vincent_hou: just one question, | 16:31 |
| vincent_hou | mtanino: Yeah. I just provide the fix today. | 16:31 |
| mtanino | vincent_hou: migrate_volume_completion() is called only the volume is not attached case, so we can solve this issue only if the volume is not attached | 16:32 |
| *** xyang1 has quit IRC | 16:32 | |
| vincent_hou | mtanino: No. The attached volume also calls this one. | 16:33 |
| vincent_hou | mtanino: It is called from nova | 16:33 |
| Swanson | model_update = {'_name_id': None} What does that mean?!? | 16:33 |
| mtanino | vincent_hou: oh, really, | 16:33 |
| vincent_hou | Yep. | 16:33 |
| *** ronis has joined #openstack-cinder | 16:34 | |
| mtanino | vincent_hou: So, if the volume path is changed suddenly after attached volume migration, I think the VM will cause I/O error because the VM does not know the renamed vg/lv, no? | 16:35 |
| *** emagana has quit IRC | 16:35 | |
| vincent_hou | Swanson: the back-end id and the volume id map to each other via the column _name_id in the table volumes. If we can update the backend id to the volume id, we do not need _name_id any more. | 16:35 |
| mtanino | vincent_hou: Let me check both nova and cinder code for attached volume migration again. | 16:36 |
| vincent_hou | mtanino: I only checked the non-attached volumes. | 16:36 |
| Swanson | vincent_hou: Thanks! I'm implementing this right now in the Dell driver. | 16:37 |
| mtanino | vincent_hou: I guess above but let me investigate the code again. | 16:37 |
| vincent_hou | mtanino: I think technically it should allow us to rename a volume attached to VMs. | 16:38 |
| vincent_hou | mtanino: I will need to verify that. | 16:38 |
| *** atan8 has quit IRC | 16:38 | |
| *** annegentle has quit IRC | 16:39 | |
| *** anuragpalsule has joined #openstack-cinder | 16:39 | |
| mtanino | vincent_hou: Doing attached volume migration using your patch is easy to understand what will happen, I will also try it. | 16:39 |
| *** rushiagr_away is now known as rushiagr | 16:40 | |
| *** rushil has quit IRC | 16:41 | |
| hemna | who is in charge of the nimblestorage CI ? | 16:41 |
| hemna | their CI seems to be broken at the moment. | 16:41 |
| *** leeantho has joined #openstack-cinder | 16:41 | |
| hemna | http://ftp.nimblestorage.com/openstack_ci/iscsi/155552/10/check/nimble-iscsi-driver-dsvm-volume/1526/logs/devstacklog.txt | 16:41 |
| *** rushil has joined #openstack-cinder | 16:42 | |
| *** emagana has joined #openstack-cinder | 16:42 | |
| *** tsekiyama has joined #openstack-cinder | 16:44 | |
| asselin | hemna, https://wiki.openstack.org/wiki/ThirdPartySystems/Nimble_Storage_CI | 16:44 |
| asselin | e-mail, but no irc contact | 16:44 |
| hemna | asselin, ok thanks | 16:44 |
| hemna | I'm just trying to be careful with the cinder os-brick patch watching all CI Failures | 16:45 |
| asselin | +1 | 16:45 |
| *** dulek has quit IRC | 16:45 | |
| *** patrickeast has joined #openstack-cinder | 16:46 | |
| *** IanGovett1 has quit IRC | 16:46 | |
| *** vincent_hou_ has joined #openstack-cinder | 16:47 | |
| hemna | ok sent them an email, asking them to look into their failures. | 16:48 |
| *** IanGovett has joined #openstack-cinder | 16:48 | |
| hemna | that failure looks like a novaclient -> keystone puke to me. | 16:48 |
| *** vincent_hou has quit IRC | 16:49 | |
| *** vincent_hou_ is now known as vincent_hou | 16:49 | |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: Remove Brick from cinder codebase https://review.openstack.org/155552 | 16:51 |
| *** swamireddy1 has joined #openstack-cinder | 16:51 | |
| *** vilobhmm1 has joined #openstack-cinder | 16:54 | |
| * e0ne bb in ~2 hours | 16:56 | |
| *** _cjones_ has quit IRC | 16:56 | |
| *** sgotliv has quit IRC | 16:59 | |
| *** Yogi11 has joined #openstack-cinder | 16:59 | |
| *** emagana has quit IRC | 17:00 | |
| *** _cjones_ has joined #openstack-cinder | 17:00 | |
| *** emagana has joined #openstack-cinder | 17:01 | |
| *** Yogi1 has quit IRC | 17:02 | |
| *** jordanP has quit IRC | 17:02 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Check volume_backend in retype https://review.openstack.org/181079 | 17:03 |
| jgriffith | smcginnis: tbarron vincent_hou ^^ | 17:03 |
| jgriffith | smcginnis: tbarron vincent_hou in my case at least that seems to fix the "orphaned" backend issue | 17:03 |
| *** Yogi11 has quit IRC | 17:04 | |
| jgriffith | pay no attention to the fact that now you're pretty well F'd when you try and figure out where the new volume is and what the mapping is | 17:04 |
| jgriffith | requires you go into the DB and manually inspect, assumign you know WTF you're even looking for | 17:04 |
| tbarron | jgriffith: in meeting but will look with interest later | 17:05 |
| *** avishay_ has joined #openstack-cinder | 17:05 | |
| *** rushil has quit IRC | 17:05 | |
| thingee | jgriffith: https://github.com/openstack/python-cinderclient/tree/1.2.1 | 17:09 |
| thingee | jgriffith: waiting for jenkins job http://logs.openstack.org/33/33a4f17fa03811a5a9d2778c8c0c8dfe343cd3bb/ | 17:10 |
| thingee | file not found... not sure if this takes a while to show up | 17:10 |
| *** Yogi1 has joined #openstack-cinder | 17:11 | |
| *** BharatK has quit IRC | 17:11 | |
| *** e0ne has quit IRC | 17:11 | |
| thingee | ah there we go | 17:12 |
| jgriffith | thingee: :) | 17:12 |
| jgriffith | thingee: patience Grass Hopper | 17:12 |
| thingee | ;) | 17:12 |
| jgriffith | thingee: that lag can be long sometimes, and annoying most times ;) | 17:13 |
| thingee | https://review.openstack.org/#/c/181083/ | 17:13 |
| thingee | jgriffith: ^ | 17:14 |
| *** vincent_hou_ has joined #openstack-cinder | 17:15 | |
| *** vincent_hou has quit IRC | 17:16 | |
| *** vincent_hou_ is now known as vincent_hou | 17:17 | |
| *** jistr has quit IRC | 17:17 | |
| *** lpetrut has quit IRC | 17:18 | |
| *** IanGovett has quit IRC | 17:19 | |
| *** garthb has quit IRC | 17:20 | |
| jgriffith | thingee: LOL "How important is workinf wifi" | 17:22 |
| jgriffith | thingee: You're a funny guy :) | 17:22 |
| vincent_hou | Have to drop off now. C U folks. | 17:22 |
| thingee | jgriffith: that's all DuncanT :) | 17:22 |
| jgriffith | vincent_hou: good night sir | 17:23 |
| *** IanGovett has joined #openstack-cinder | 17:23 | |
| *** vincent_hou has quit IRC | 17:24 | |
| *** xyang1 has joined #openstack-cinder | 17:26 | |
| *** merooney has quit IRC | 17:33 | |
| *** merooney has joined #openstack-cinder | 17:35 | |
| *** ctina__ has quit IRC | 17:37 | |
| *** ctina___ has joined #openstack-cinder | 17:37 | |
| *** emagana has quit IRC | 17:40 | |
| *** annegentle has joined #openstack-cinder | 17:40 | |
| *** erlon has joined #openstack-cinder | 17:43 | |
| *** emagana has joined #openstack-cinder | 17:43 | |
| *** avishay_ has quit IRC | 17:44 | |
| *** koralgolek has joined #openstack-cinder | 17:45 | |
| *** anteaya has quit IRC | 17:45 | |
| koralgolek | Hi everyone, I start using openstack and I'm wondering if is it possible to attach directly local disk to vm? I guess cinder throught iscsi would be slower than local storage. Local storage I also mean LUN by Fibre Channel. | 17:45 |
| *** annegentle has quit IRC | 17:46 | |
| *** swamireddy1 has quit IRC | 17:48 | |
| *** emagana has quit IRC | 17:48 | |
| *** kaisers has quit IRC | 17:48 | |
| jgriffith | koralgolek: there's not a built in mechanism for what you're describing | 17:49 |
| *** jistr has joined #openstack-cinder | 17:49 | |
| *** kaisers has joined #openstack-cinder | 17:49 | |
| jgriffith | koralgolek: although you contradict your self a bit there "Local storage I also mean LUN by Fibre Channel." | 17:49 |
| jgriffith | so I guess the answer techinally is "yes" and "no" at the same time | 17:50 |
| *** angela-s has joined #openstack-cinder | 17:50 | |
| patrickeast | ^ maybe there is already a FC LUN on the compute node and they want to use it for ephemeral storage? | 17:50 |
| jgriffith | patrickeast: ephemeral != persistent | 17:50 |
| *** madskier has joined #openstack-cinder | 17:51 | |
| patrickeast | jgriffith: well yea, they never said persistent | 17:51 |
| patrickeast | just local storage | 17:51 |
| jgriffith | patrickeast: but regardless... san/FC is NOT local disk either | 17:51 |
| patrickeast | haha true | 17:51 |
| koralgolek | The problem is that I've several servers with hba cards yet. im using now ganeti and to whole servers there is presented LUN with LVS VG | 17:51 |
| jgriffith | that was actually my main point there | 17:51 |
| jgriffith | and perhaps an FC device with a cinder FC driver | 17:52 |
| koralgolek | every lv is separate vm | 17:52 |
| jgriffith | koralgolek: yeah, sorry I'm not going to be much help | 17:52 |
| *** theanalyst has quit IRC | 17:52 | |
| koralgolek | "patrickeast: but regardless... san/FC is NOT local disk either" I know its not but on host server it is visible as block storage | 17:54 |
| *** theanalyst has joined #openstack-cinder | 17:55 | |
| *** merooney has quit IRC | 17:56 | |
| *** garthb has joined #openstack-cinder | 17:58 | |
| *** rushiagr is now known as rushiagr_away | 18:00 | |
| koralgolek | btw assume Im gonna use classic cinder. Is there possible to setup few nodes of it to make load balance? | 18:02 |
| jgriffith | koralgolek: sorry, I don't know what "classic" cinder is | 18:03 |
| koralgolek | every node will have the same lun with the same lvm - maybe even clvm | 18:03 |
| *** vilobhmm1 has quit IRC | 18:03 | |
| *** vilobhmm1 has joined #openstack-cinder | 18:03 | |
| jgriffith | koralgolek: clustered LVM type thing.. mtanino I believe had some proposals around thetype of thing you're describing | 18:03 |
| *** vilobhmm1 has quit IRC | 18:04 | |
| jgriffith | koralgolek: at least the multi-attach LVM thing | 18:04 |
| *** vilobhmm1 has joined #openstack-cinder | 18:04 | |
| *** emagana has joined #openstack-cinder | 18:05 | |
| *** annegentle has joined #openstack-cinder | 18:06 | |
| *** e0ne has joined #openstack-cinder | 18:06 | |
| *** vmtrooper has joined #openstack-cinder | 18:07 | |
| *** _cjones_ has quit IRC | 18:07 | |
| koralgolek | classic cinder I mean machine with persistent or not persistent sotarage with lvm which is presented to computed nodes throught iscsi (am I right?) | 18:08 |
| hemna | looks like the microsoft CI is failing due to a local.conf REQUIREMENTS_BRANCH setting. | 18:08 |
| hemna | http://64.119.130.115/cinder/155552/11/iscsi/devstack_config/local.conf.gz | 18:09 |
| hemna | not sure why they have that in there. | 18:09 |
| hemna | it will fail to pickup any changes in the global requirements repo | 18:09 |
| hemna | which is what it's doing for os-brick | 18:09 |
| hemna | :( | 18:09 |
| hemna | + python update.py /opt/stack/cinder | 18:09 |
| hemna | 'os-brick' is not in global-requirements.txt | 18:09 |
| jgriffith | hemna: I suspect that's why they're doing it | 18:09 |
| hemna | when it's there. | 18:09 |
| *** _cjones_ has joined #openstack-cinder | 18:09 | |
| hemna | I'll email them | 18:10 |
| *** dencaval has quit IRC | 18:10 | |
| jgriffith | hemna: https://review.openstack.org/#/c/171060/ | 18:11 |
| *** vmtrooper has quit IRC | 18:11 | |
| hemna | hrmm | 18:12 |
| hemna | maybe I'm wrong, but doesn't local.conf REQUIREMENTS_BRANCH force a specific branch of the requirements repo ? which I think is missing the os-brick patch that just merged | 18:12 |
| *** jordanP has joined #openstack-cinder | 18:12 | |
| jgriffith | hemna: FWIW, you might have better luck adding the requirement "before" making the switch-over | 18:12 |
| hemna | so a separate patch just to add os-brick to cinder's requirements.txt ? | 18:13 |
| *** jordanP has quit IRC | 18:13 | |
| hemna | then wait a few hours kinda deal ? | 18:13 |
| jgriffith | hemna: yeah... well, that part sucks :( | 18:13 |
| jgriffith | hemna: but you can make your next series a dep | 18:13 |
| hemna | heh ok. I'll try that. | 18:13 |
| *** vilobhmm1 has quit IRC | 18:13 | |
| jgriffith | hemna: and to your other question "yes" | 18:14 |
| jgriffith | hemna: the thing is it appears the microsoft team "needed" that installer patch to work properly | 18:14 |
| jgriffith | hemna: so they froze on that branch | 18:14 |
| jgriffith | hemna: now that the change has merged they should be able to remove it | 18:14 |
| hemna | probably why they pinned it to that | 18:14 |
| jgriffith | alexpilotti: ^^ | 18:15 |
| *** merooney has joined #openstack-cinder | 18:15 | |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: Add os-brick to cinder requirements.txt https://review.openstack.org/181102 | 18:16 |
| *** anuragpalsule has quit IRC | 18:16 | |
| *** vilobhmm1 has joined #openstack-cinder | 18:18 | |
| *** adam_g has left #openstack-cinder | 18:18 | |
| *** anteaya has joined #openstack-cinder | 18:18 | |
| *** ctina___ has quit IRC | 18:20 | |
| *** merooney has quit IRC | 18:20 | |
| *** vilobhmm1 has quit IRC | 18:21 | |
| *** vilobhmm1 has joined #openstack-cinder | 18:22 | |
| tbarron | jgriffith: w.r.t. https://review.openstack.org/#/c/181079/ do you need to make a similar check for the fallback migrate path? (I could be missing something fundamental here, dunno much about migration.) | 18:22 |
| jgriffith | tbarron: well, that's *why* the fall-back migrate path is there | 18:23 |
| jgriffith | tbarron: if the driver.retype fails or can't be done, it does the migrate process | 18:23 |
| jgriffith | tbarron: the whole point of migrate in that context is that the types aren't compatible on the same backend... see what i mean? | 18:23 |
| *** vilobhmm1 has quit IRC | 18:24 | |
| mtanino | koralgolek: default Cinder LVM driver provides LV as a volume via iISCSI using tgtd, LIO, iet etc, | 18:24 |
| *** vilobhmm1 has joined #openstack-cinder | 18:24 | |
| tbarron | kk, that's what I was suspecting. | 18:24 |
| *** ebalduf has quit IRC | 18:25 | |
| tbarron | jgriffith: a pain, but you should modify the retype tests in unit/test_volume.py, no? | 18:25 |
| tbarron | looks like they just test when extra-specs are {} for old and new at a quick glance | 18:26 |
| *** madskier has quit IRC | 18:26 | |
| jgriffith | tbarron: I'm pretty opposed to doing things like writing unit tests to see if an "if" statement works | 18:27 |
| *** MentalRay has quit IRC | 18:27 | |
| hemna | hrmm guess you can't email the microsoft cinder ci email address | 18:27 |
| hemna | it bounced | 18:27 |
| hemna | heh | 18:27 |
| tbarron | jgriffith: well, I'm not trying to start a unit-test holy ware atm | 18:27 |
| tbarron | s/ware/war/ | 18:28 |
| jgriffith | tbarron: me neither :) | 18:28 |
| smcginnis | hemna: You have to send it using Outlook. :) | 18:28 |
| hemna | gah | 18:28 |
| *** ebalduf has joined #openstack-cinder | 18:28 | |
| jgriffith | hemna: you're not fooling anyone, we know you have Outlook | 18:28 |
| hemna | :P | 18:29 |
| smcginnis | He just uses linux for show. He has a whole active directory, Exchange home network set up. | 18:29 |
| hemna | sssh | 18:29 |
| hemna | smcginnis, no beer for you! | 18:29 |
| *** merooney has joined #openstack-cinder | 18:30 | |
| smcginnis | hehhe | 18:30 |
| tbarron | jgriffith: will just say, before ducking, that the point of the unit test is not to verify that the if works, but rather that someone doesn't break your fix later on. | 18:30 |
| * tbarron ducks now | 18:30 | |
| koralgolek | mtanino: to be sure - so there is NO possibility to serve storage directly from FC LUN or local distk to vm on given host server? | 18:30 |
| jgriffith | tbarron: LOL, that's really not true actually. If someobdy makes a code change and it breaks the test 9.9 times out of 10, they'll just modify the test to work :) | 18:31 |
| openstackgerrit | Vilobh Meshram proposed openstack/cinder: Bad link in API version details response https://review.openstack.org/180691 | 18:31 |
| mtanino | koralgolek: If the local means internal HDD, the only way to use the HDD is using LVM driver and pass it via iSCSI using tgtd. | 18:34 |
| *** eharney_ has joined #openstack-cinder | 18:35 | |
| mtanino | koralgolek: but if your FC storage has Cinder driver, you can attach a LUN as a volume to your VM directly(via FC/iSCSI) using that Cinder driver. | 18:35 |
| koralgolek | mtanino: then I need to install cinder on every node? | 18:36 |
| mtanino | koralgolek: no, it is enough to install cinder volume, scheduler, api only for one node. | 18:37 |
| *** eharney has quit IRC | 18:38 | |
| mtanino | koralgolek: but you need to install nova compute to other hyper-visor nodes. | 18:38 |
| koralgolek | mtanino: sorry maybe for stupid question but why do I need Cinder driver to use FC LUN when on servers I've already configured multipathd and I see LUNs like block devices? | 18:38 |
| hemna | koralgolek, guests in kvm don't have an FC HBA | 18:39 |
| hemna | only the host OS has the FC HBA available | 18:39 |
| *** timcl has quit IRC | 18:40 | |
| hemna | unfortunately, libvirt can't do a virtual FC HBA PCI device :( | 18:40 |
| koralgolek | I know but Im wondering why can't I attach block device directly to vm? | 18:40 |
| hemna | the block device has to be attached to the host first | 18:40 |
| koralgolek | for example I can do pvcreate, vgcreate and on this finally lvcreate | 18:40 |
| hemna | then passed in to the guest | 18:40 |
| hemna | because of what I just explained. | 18:40 |
| hemna | if it's an FC LUN | 18:41 |
| jgriffith | hemna: koralgolek time out :) | 18:41 |
| jgriffith | koralgolek: so the reason you "can't" currently is only because that's not how Cinder was designed | 18:41 |
| * hemna sits in the corner | 18:41 | |
| hemna | :P | 18:41 |
| mtanino | koralgolek: so the answer is Cinder is an abstracted storage management layer I think. | 18:42 |
| jgriffith | koralgolek: what you want to do is completely "possible" with KVM (local disk already attached and just pass through) just like we do today | 18:42 |
| jgriffith | mtanino: +1 | 18:42 |
| jgriffith | koralgolek: there's an LVM driver and components in Nova | 18:42 |
| jgriffith | koralgolek: you could certainly tweak some things around to do what you want | 18:42 |
| jgriffith | koralgolek: you could even manipulate Cinder to manage/orchestrate some of it | 18:42 |
| jgriffith | koralgolek: but it's not very straight forward | 18:43 |
| jgriffith | koralgolek: and you probably won't get much help in the dev channel | 18:43 |
| koralgolek | :( | 18:43 |
| jgriffith | hemna: so you're correct, but keep in mind his proposal is similar to what was discussed by Hitachi in Paris | 18:43 |
| hemna | oh, another one of those | 18:43 |
| jgriffith | hemna: you maually attach FC luns to all the compute notes, make them all the same shared VG | 18:44 |
| *** eharney_ is now known as eharney | 18:44 | |
| jgriffith | and carve LV's out of them. That means "all" compute nodes are sharing the VG/LV's | 18:44 |
| koralgolek | excatly!! | 18:44 |
| jgriffith | koralgolek: yeah, I get it... I was just avoiding it :) | 18:44 |
| koralgolek | I use this setup on ganeti | 18:44 |
| jgriffith | koralgolek: the fact is that I'm pretty opposed to that concept | 18:45 |
| hemna | I'm not sure why you would ever want to do that | 18:45 |
| hemna | but to each their own I suppose. | 18:45 |
| jgriffith | hemna: there are lots of good reasons actually | 18:45 |
| jgriffith | hemna: you don't need a driver in Cinder being one of the best | 18:45 |
| hemna | well I'm not sure that's a good reason :) | 18:45 |
| jgriffith | hemna: the other reason is it makes things like multi-attach obsolete and irrelevant | 18:46 |
| jgriffith | hemna: you get it for free | 18:46 |
| koralgolek | in my opinion the main reason is performance | 18:46 |
| jgriffith | hemna: you just flip a bit in Cinder to turn it on | 18:46 |
| hemna | well for linux based hypervisors sure. | 18:46 |
| jgriffith | koralgolek: now that's where I disagree :) | 18:46 |
| jgriffith | koralgolek: I benchmarked this when it came up a while back | 18:47 |
| hemna | performance is not something to highlight in this one | 18:47 |
| jgriffith | koralgolek: I saw ZERO improvement in performance | 18:47 |
| hemna | it's not for performance at all. IMHO | 18:47 |
| jgriffith | koralgolek: the only performance tests that were shown that had a significant improvement did things like: | 18:47 |
| *** JayJ_ has quit IRC | 18:47 | |
| hemna | adding LVM on top of an FC VOLUME makes for better performance ?!?!? | 18:47 |
| koralgolek | jgrffith: even in number of IOs? | 18:48 |
| koralgolek | per second | 18:48 |
| jgriffith | compare FC attached luns to iSCSI over 1Gig | 18:48 |
| jgriffith | koralgolek: if comparing apples to apples I wasn't able to see a different | 18:48 |
| jgriffith | difference | 18:48 |
| jgriffith | DuncanT: however stated he tested it and saw a sifnificant difference | 18:48 |
| *** reed has quit IRC | 18:49 | |
| jgriffith | I went home that night and tried it again but couldn't arrive at the same results.. dont' know why | 18:49 |
| jgriffith | koralgolek: in most cases the transport layer isn't the bottle-neck for performance these days (in most cases) | 18:49 |
| DuncanT | jgriffith: It depends heavily on what you're using as your c-col node, it turns out | 18:49 |
| jgriffith | koralgolek: I'll put 10G iSCSI against FC and sometimes win | 18:49 |
| jgriffith | sometimes loose, but it's rarely because of the transport layer | 18:50 |
| jgriffith | DuncanT: now that makes sense of course | 18:50 |
| jgriffith | DuncanT: because of the tgt work and CPU/Memory load it takes | 18:50 |
| koralgolek | correct me if im wrong but I dont belevie one thig. I have 20 servers with many vm's and I dont believe that one target node (cinder) will manage these all servers | 18:50 |
| jgriffith | DuncanT: you have a big enough c-vol node it doesn't matter | 18:51 |
| *** jistr has quit IRC | 18:51 | |
| jgriffith | DuncanT: in my case/testing I used SolidFire so that part wasn't a factor | 18:51 |
| DuncanT | jgriffith: And how fast you backend is.... fast backend and a relatively weedy c-vol node (or a crappy 10 gig card, which I think was the bottleneck for us... drive the 10gig hard and the CPU goes nuts) and you loose big time | 18:51 |
| jgriffith | again apples to apples | 18:51 |
| jgriffith | DuncanT: yeah, that's fair | 18:51 |
| DuncanT | thangp: You about? | 18:51 |
| *** MentalRay has joined #openstack-cinder | 18:52 | |
| koralgolek | what about IOPS in this case? | 18:52 |
| jgriffith | koralgolek: that's the only real test of performance IMO | 18:52 |
| thangp | DuncanT: hey | 18:52 |
| jgriffith | koralgolek: anyway.. I gotta run. Good luck to ya | 18:52 |
| koralgolek | thx jgriffith ;) | 18:53 |
| thangp | DuncanT: ready to talk about the rpc version compat? | 18:53 |
| *** emagana has quit IRC | 18:53 | |
| DuncanT | thangp: Yup | 18:53 |
| thangp | DuncanT: :) cool, what are your plans? | 18:53 |
| thangp | DuncanT: btw, i was able to prototype a way to do object version compat: https://review.openstack.org/#/c/180393/ | 18:54 |
| thangp | DuncanT: I tested it on a 2015.1.1 and 2015.2.0 release | 18:54 |
| *** dencaval has joined #openstack-cinder | 18:54 | |
| *** emagana has joined #openstack-cinder | 18:55 | |
| *** emagana has quit IRC | 18:55 | |
| openstackgerrit | Walter A. Boring IV (hemna) proposed openstack/cinder: Add os-brick to cinder requirements.txt https://review.openstack.org/181102 | 18:56 |
| DuncanT | thangp: Cool, I'll take a look at that in a sec | 18:56 |
| *** reed has joined #openstack-cinder | 18:57 | |
| DuncanT | Anybody having issues with etherpad? | 18:58 |
| DuncanT | I keep getting disconnected | 18:58 |
| *** emagana has joined #openstack-cinder | 18:59 | |
| vilobhmm1 | thingee : ping | 18:59 |
| DuncanT | thangp: Trying to type up a few thoughts on an etherpad | 18:59 |
| thingee | vilobhmm1: hi | 18:59 |
| DuncanT | https://etherpad.openstack.org/p/duncant-on-versioned-things | 19:00 |
| vilobhmm1 | can you please have a look at https://review.openstack.org/#/c/173141/ | 19:00 |
| thangp | DuncanT: ok, np | 19:00 |
| vilobhmm1 | seems ok for me DuncanT | 19:00 |
| e0ne | can't find myself. will we remove APIv1 from Cinder in Liberty | 19:00 |
| e0ne | ? | 19:00 |
| vilobhmm1 | thingee : https://review.openstack.org/#/c/173141/ seems to have got blessing from most of the members of our team | 19:00 |
| thangp | DuncanT: I got in | 19:00 |
| thingee | vilobhmm1: great | 19:01 |
| vilobhmm1 | if you also find it ok i can officially start posting changes for review | 19:01 |
| thingee | vilobhmm1: yes that would be the plan. | 19:01 |
| vilobhmm1 | thingee : great…so please have a look when you get time… | 19:02 |
| vilobhmm1 | thnx | 19:02 |
| DuncanT | Grrrr, infernal network connection! | 19:02 |
| thangp | DuncanT: google docs? | 19:02 |
| *** MentalRay has quit IRC | 19:02 | |
| thangp | DuncanT: another alt | 19:02 |
| thingee | vilobhmm1: I'm aware of specs pending for review, thank you. | 19:02 |
| vilobhmm1 | thingee : cool | 19:03 |
| vilobhmm1 | ok | 19:03 |
| *** merooney has quit IRC | 19:04 | |
| vilobhmm1 | DuncanT : regarding our talk apart from the points mentioned in the etherpad do you want to cover anything else like interaction with nova, using compare and set (at the db layer) | 19:04 |
| vilobhmm1 | right now also i feel we have enough to cover but i think it would be nice to cover things like using compare and set for guaranting atomicity | 19:06 |
| *** merooney has joined #openstack-cinder | 19:06 | |
| DuncanT | vilobhmm1: I'll get onto that once I've finished writing up my thoughts on versioning. Things like compare and set .v. DLM need to be talked about... the benchmarks were not promising | 19:06 |
| vilobhmm1 | sure… | 19:06 |
| vilobhmm1 | please ping me when you are free | 19:06 |
| DuncanT | compare and set (update status='new' where id='xxxyyyzzz' and status='oldstatus') looked best in benchmarks | 19:06 |
| vilobhmm1 | yup…thats the way we should go… | 19:07 |
| DuncanT | thangp: Typing up elsewhere and will paste in a minute | 19:09 |
| *** fthiagogv has quit IRC | 19:09 | |
| thangp | DuncanT: ok np | 19:09 |
| vilobhmm1 | is anyone aware from where does Cinder API docs get generated ? does it get built by manuals or something in cinder codebase invokes it | 19:11 |
| annegentle | vilobhmm1: it's in the openstack/api-site repo | 19:11 |
| annegentle | vilobhmm1: in WADL, handwritten, maintained | 19:11 |
| vilobhmm1 | annegentle : cool thnx | 19:11 |
| annegentle | vilobhmm1: I'm working on a spec to start automating just the ref info | 19:11 |
| annegentle | vilobhmm1: what's your interest area? | 19:11 |
| vilobhmm1 | we have some unit test for them in https://github.com/openstack/cinder/tree/master/cinder/tests/unit/api | 19:12 |
| vilobhmm1 | and i was working on on bug where the links were broken | 19:12 |
| vilobhmm1 | i checked openstack-manuals and cinder codebase | 19:12 |
| annegentle | vilobhmm1: last release we did move an openstack/volume-api repo to openstack-attic/volume-api | 19:12 |
| annegentle | vilobhmm1: and some of the info went to api-site and some went to cinder-specs | 19:13 |
| annegentle | vilobhmm1: http://specs.openstack.org/openstack/cinder-specs/#volume-v2-api | 19:13 |
| annegentle | vilobhmm1: http://developer.openstack.org/api-ref-blockstorage-v2.html | 19:13 |
| vilobhmm1 | but couldn't find one so..was curious to know…thanks for the details annegentle... | 19:13 |
| annegentle | those are the links that are now official | 19:13 |
| annegentle | vilobhmm1: here to serve | 19:14 |
| vilobhmm1 | annegentle : i would like you to review https://review.openstack.org/#/c/180691/ | 19:14 |
| annegentle | vilobhmm1: looking now | 19:14 |
| vilobhmm1 | cool | 19:14 |
| vilobhmm1 | thanks | 19:14 |
| openstackgerrit | Mike Perez proposed openstack/cinder: Allow rexports for targets with Datera https://review.openstack.org/181120 | 19:15 |
| openstackgerrit | Mike Perez proposed openstack/cinder: Rename Datera test to test_datera for discovery https://review.openstack.org/181121 | 19:15 |
| annegentle | vilobhmm1: yep that's the PDF you want | 19:15 |
| vilobhmm1 | awesome | 19:15 |
| annegentle | vilobhmm1: oh actually, no. that's the v1, you want v2.0 based on the ID in line 91? | 19:16 |
| *** nikeshm has joined #openstack-cinder | 19:16 | |
| nikeshm | patrickeast asselin : hi | 19:16 |
| vilobhmm1 | sure you can comment that on the review will be easy | 19:16 |
| patrickeast | nikeshm: hey | 19:16 |
| vilobhmm1 | annegentle : ^^ | 19:16 |
| nikeshm | is supports for ubuntu 14.04 | 19:17 |
| nikeshm | CI | 19:17 |
| patrickeast | nikeshm: for which part? | 19:17 |
| *** merooney has quit IRC | 19:17 | |
| nikeshm | patrickeast:jenkins master | 19:17 |
| patrickeast | nikeshm: yea i use 14.04 for my masters | 19:17 |
| patrickeast | nikeshm: and for the nodepool nodes | 19:18 |
| annegentle | vilobhmm1: yep, done | 19:18 |
| nikeshm | patrickeast : where nodepool comes in picture in jenkins,zuul and gearman | 19:18 |
| vilobhmm1 | annegentle : thanks | 19:19 |
| annegentle | vilobhmm1: thank You! | 19:19 |
| *** IlyaG has quit IRC | 19:20 | |
| patrickeast | nikeshm: nodepool runs in the background and provides workers for jenkins to run jobs on | 19:21 |
| *** _cjones_ has quit IRC | 19:21 | |
| patrickeast | nikeshm: the zuul->gearman->jenkins flow works the same with our without nodepool (for the most part) | 19:22 |
| nikeshm | nikesm: still i was facing issue with image-build,ok will try again,BTW,when i am trying to upload a patch in sandbox,i am able to see noopcheck in build history of jenkins saying (pending—Waiting for next available executor) | 19:23 |
| nikeshm | its not able to build | 19:24 |
| patrickeast | nikeshm: its probably waiting for a build node with the correct label, if nodepool hasn’t been able to build an image and attach slaves to jenkins they won’t be able to run | 19:24 |
| nikeshm | pending from long time | 19:24 |
| nikeshm | ohhh | 19:25 |
| *** madskier has joined #openstack-cinder | 19:25 | |
| nikeshm | is this nodepool yaml file http://paste.openstack.org/show/216406/ | 19:27 |
| nikeshm | patrickeast: is this nodepool yaml file http://paste.openstack.org/show/216406/ do we have to create a image 'dpc' in glance manually | 19:28 |
| *** abhiram_moturi has joined #openstack-cinder | 19:28 | |
| DuncanT | thangp: Hopefully what is up now makes some sense? | 19:28 |
| nikeshm | patrickeast : is this yaml fine? do we have do create a image manually in glance | 19:28 |
| thangp | DuncanT: so my question is - why not have some kind of rpc function to grab the version? | 19:29 |
| DuncanT | thangp: How do you version that RPC? Who do you send it to? | 19:30 |
| *** madskier has quit IRC | 19:30 | |
| thangp | DuncanT: we could have a rpc function that never changes, so no need to version... | 19:31 |
| thangp | e.g. get_version | 19:31 |
| DuncanT | thangp: Ok, who sends it to who, and when? | 19:31 |
| DuncanT | thangp: The service coming up has to send it, since only it knows it just came up | 19:32 |
| *** merooney has joined #openstack-cinder | 19:32 | |
| thangp | DuncanT: brb...got a quick mtg | 19:32 |
| DuncanT | thangp: Ok, I should be here for a while | 19:32 |
| *** bkopilov has quit IRC | 19:33 | |
| thangp | DuncanT: mtg canceled, so i'm back | 19:33 |
| DuncanT | Best kind of meeting | 19:35 |
| thangp | DuncanT: so usually a host is selected before an rpc call is made, right? e.g. new_host = utils.extract_host(volume['host']) | 19:35 |
| DuncanT | thangp: Ok... | 19:36 |
| DuncanT | thangp: Though sheduler is bcast | 19:36 |
| thangp | DuncanT: since we know the host, we can ask the version, make the rpc version and obj version compatibile, then send? | 19:36 |
| thangp | DuncanT: I have to look at scheduler | 19:36 |
| thangp | DuncanT: not sure | 19:37 |
| DuncanT | thangp: That works for c-vol but not sched and not backup in future | 19:37 |
| DuncanT | It also means every RPC becomes a syncronous RPC + a second fire and forget | 19:37 |
| thangp | DuncanT: schedule does a cast so yes, its a broadcast | 19:38 |
| DuncanT | So if a c-vol process is hung, then you start eating up workers on API/scheduler, waiting for it to respond | 19:38 |
| DuncanT | as well as being slower | 19:38 |
| DuncanT | even when everything works | 19:38 |
| *** ociuhandu has quit IRC | 19:38 | |
| thangp | DuncanT: ok, i get what you're saying | 19:39 |
| DuncanT | thangp: Ideally I want services to suicide before connecting to the RPC bus if they don't support the minimum version | 19:40 |
| DuncanT | thangp: And the advantage of the DB approach is it can all be cached at startup, and no extra work is needed until the get an upgrade message | 19:41 |
| thangp | DuncanT: reading etherpad... | 19:42 |
| *** bkopilov has joined #openstack-cinder | 19:43 | |
| *** timcl has joined #openstack-cinder | 19:44 | |
| *** abhiram_moturi has quit IRC | 19:46 | |
| openstackgerrit | Rajini Ram proposed openstack/cinder: Eqlx: Fixes the retries on Network Connection Error https://review.openstack.org/171736 | 19:46 |
| *** abhiram_moturi has joined #openstack-cinder | 19:46 | |
| *** merooney has quit IRC | 19:48 | |
| thangp | DuncanT: with your plan, how do we know when all the services are upgraded? | 19:48 |
| *** MentalRay has joined #openstack-cinder | 19:48 | |
| thangp | DuncanT: so that the services will use the new rpc/obj versions instead of the pinned | 19:48 |
| nikeshm | patrickeast: there? | 19:49 |
| DuncanT | thangp: They're running the new version of the software when they've written in the new, higher max version at startup. They're using the new rpc/obj version when they've cleared the desired field and updated the pin | 19:49 |
| *** IlyaG has joined #openstack-cinder | 19:49 | |
| *** bkopilov has quit IRC | 19:53 | |
| thangp | DuncanT: so when a rpc call is made or an obj is instantiated, we have to check the db for the version to make compatible to? | 19:53 |
| *** bkopilov has joined #openstack-cinder | 19:53 | |
| DuncanT | thangp: I suggest it is read at startup and cached in a global | 19:53 |
| DuncanT | thangp: It can't change after startup until an 'upgrade your version now' rpc is received | 19:54 |
| thangp | DuncanT: ah ok | 19:55 |
| *** ronis has quit IRC | 19:55 | |
| thangp | so who ends up sending the 'upgrade your version'? | 19:55 |
| *** ronis has joined #openstack-cinder | 19:55 | |
| DuncanT | The API service will send it to all 'up' services, in response to an admin request (i.e. it isn't automatic) | 19:56 |
| *** merooney has joined #openstack-cinder | 19:56 | |
| thangp | ok, so to summarize this based on my understanding... | 19:56 |
| thangp | each service gets updated individually, when it is started, it updates the db with its supported version. in the end, when all services are updated, the admin sends an update request to use the latest version | 19:58 |
| DuncanT | thangp: Yup | 19:58 |
| thangp | DuncanT: got it now :) | 19:59 |
| *** bkopilov has quit IRC | 19:59 | |
| thangp | DuncanT: sounds like a good plan to me | 19:59 |
| *** bkopilov has joined #openstack-cinder | 19:59 | |
| *** delattec has quit IRC | 19:59 | |
| *** annegentle has quit IRC | 20:02 | |
| *** annegentle has joined #openstack-cinder | 20:02 | |
| *** koralgolek has quit IRC | 20:02 | |
| *** _cjones_ has joined #openstack-cinder | 20:03 | |
| *** dencaval has quit IRC | 20:03 | |
| DuncanT | thangp: Awesome :-) | 20:03 |
| vilobhmm1 | DuncanT : but cinder.services doesn't not contain any version info | 20:03 |
| DuncanT | vilobhmm1: Yet | 20:04 |
| vilobhmm1 | the version_string is just computed here https://github.com/openstack/cinder/blob/master/cinder/service.py#L140 | 20:04 |
| vilobhmm1 | right | 20:04 |
| DuncanT | thangp: Please feel free to add / edit the etherpad | 20:04 |
| thangp | thangp: ok | 20:04 |
| vilobhmm1 | so in the above summary that thang mentioned how does version get updated to db ? | 20:04 |
| *** MentalRay has quit IRC | 20:05 | |
| thangp | vilobhmm1: we have to add to start() to do that | 20:05 |
| vilobhmm1 | i know so my question is will there be a new field called as "version" in cinder.services ? | 20:07 |
| vilobhmm1 | or for that matter in any other table ? | 20:07 |
| DuncanT | vilobhmm1: Yes. Or maybe one object_version and one rpc_version | 20:08 |
| *** alexpilotti has quit IRC | 20:08 | |
| vilobhmm1 | DuncanT : now it makes sense | 20:08 |
| thangp | DuncanT: we need a min and max version, no? | 20:08 |
| DuncanT | thangp: Yeah, or 'supported' and 'pinned' might be better (clearer) names | 20:09 |
| *** alexpilotti has joined #openstack-cinder | 20:09 | |
| thangp | DuncanT: ok | 20:09 |
| vilobhmm1 | sounds good | 20:09 |
| asselin | nikeshm, hi | 20:10 |
| *** rushil has joined #openstack-cinder | 20:12 | |
| *** jkremer has joined #openstack-cinder | 20:14 | |
| *** MentalRay has joined #openstack-cinder | 20:15 | |
| *** alexpilotti has quit IRC | 20:16 | |
| *** breitz has quit IRC | 20:17 | |
| *** breitz has joined #openstack-cinder | 20:17 | |
| *** Longgeek has quit IRC | 20:18 | |
| *** AlkaD has joined #openstack-cinder | 20:20 | |
| *** _cjones_ has quit IRC | 20:21 | |
| *** alexpilotti has joined #openstack-cinder | 20:21 | |
| hemna | https://review.openstack.org/#/c/181102/ | 20:23 |
| patrickeast | nikeshm: sry, stepped out of the office for a little bit, that image should be re-built from that cron setting, so like i’ve configured mine to a new one every day at 2am, you shouldn’t have to upload it manually to glance, nodepool does it for you | 20:25 |
| *** timcl has quit IRC | 20:25 | |
| *** _cjones_ has joined #openstack-cinder | 20:25 | |
| *** drjones has joined #openstack-cinder | 20:25 | |
| *** abhiram_moturi has quit IRC | 20:27 | |
| *** abhiram_moturi has joined #openstack-cinder | 20:28 | |
| *** ronis has quit IRC | 20:28 | |
| *** alexpilotti has quit IRC | 20:29 | |
| *** ebalduf has quit IRC | 20:30 | |
| *** emagana has quit IRC | 20:31 | |
| *** ebalduf has joined #openstack-cinder | 20:31 | |
| asselin | nikeshm, I still recomend manually building the node image the first time: nodepool image-build dpc (as nodepool user) | 20:32 |
| *** emagana has joined #openstack-cinder | 20:32 | |
| openstackgerrit | Derrick Wippler proposed openstack/python-cinderclient: Avoid _get_keystone_session() if auth_plugin https://review.openstack.org/155518 | 20:34 |
| *** bnemec has quit IRC | 20:36 | |
| *** abhiram_moturi has quit IRC | 20:41 | |
| e0ne | hemna: why do we need separate patch to add a dependency? | 20:41 |
| *** abhiram_moturi has joined #openstack-cinder | 20:42 | |
| *** abhiram_moturi has quit IRC | 20:44 | |
| *** abhiram_moturi has joined #openstack-cinder | 20:44 | |
| hemna | e0ne, mirror delays | 20:44 |
| e0ne | hemna: thanks! didn't know about such hack:) | 20:46 |
| * e0ne bb in ~12 hours | 20:46 | |
| *** abhiram_moturi has quit IRC | 20:48 | |
| *** abhiram_moturi has joined #openstack-cinder | 20:49 | |
| *** abhiram_moturi has quit IRC | 20:51 | |
| *** lpabon has quit IRC | 20:51 | |
| *** abhiram_moturi has joined #openstack-cinder | 20:51 | |
| *** e0ne has quit IRC | 20:51 | |
| *** bnemec has joined #openstack-cinder | 20:51 | |
| *** abhiram_moturi has quit IRC | 20:53 | |
| asselin | nikeshm, also export DIB_OFFLINE=true to make it go faster | 20:53 |
| *** abhiram_moturi has joined #openstack-cinder | 20:54 | |
| openstackgerrit | Derrick Wippler proposed openstack/python-cinderclient: Avoid _get_keystone_session() if auth_plugin https://review.openstack.org/155518 | 20:54 |
| *** bswartz has quit IRC | 20:57 | |
| *** akerr has quit IRC | 20:58 | |
| *** jungleboyj has quit IRC | 21:00 | |
| *** merooney has quit IRC | 21:02 | |
| *** thangp has quit IRC | 21:04 | |
| *** annegentle has quit IRC | 21:04 | |
| *** abhiram_moturi has quit IRC | 21:06 | |
| *** abhiram_moturi has joined #openstack-cinder | 21:06 | |
| *** vmtrooper has joined #openstack-cinder | 21:08 | |
| *** Lee1092 has quit IRC | 21:11 | |
| *** juzuluag has quit IRC | 21:12 | |
| *** Anticime1 is now known as Anticimex | 21:14 | |
| *** EmilienM is now known as EmilienM|afk | 21:16 | |
| *** abhiram_moturi has quit IRC | 21:18 | |
| *** abhiram_moturi has joined #openstack-cinder | 21:18 | |
| *** abhiram_moturi has quit IRC | 21:20 | |
| *** abhiram_moturi has joined #openstack-cinder | 21:21 | |
| *** bnemec has quit IRC | 21:25 | |
| *** IlyaG has quit IRC | 21:28 | |
| *** Yogi1 has quit IRC | 21:28 | |
| *** IlyaG has joined #openstack-cinder | 21:31 | |
| *** IanGovett has quit IRC | 21:33 | |
| openstackgerrit | Matt Riedemann proposed openstack/cinder: Log command failure details before raising ISCSITargetDetachFailed https://review.openstack.org/181157 | 21:35 |
| *** akerr has joined #openstack-cinder | 21:36 | |
| *** akerr_ has joined #openstack-cinder | 21:37 | |
| *** jkremer has quit IRC | 21:37 | |
| *** akerr has quit IRC | 21:41 | |
| *** IlyaG has quit IRC | 21:42 | |
| openstackgerrit | Tom Swanson proposed openstack/cinder: Dell: Added support for update_migrated_volume https://review.openstack.org/181160 | 21:44 |
| *** bill_az has joined #openstack-cinder | 21:47 | |
| *** dustins has quit IRC | 21:49 | |
| *** dustins has joined #openstack-cinder | 21:50 | |
| *** IlyaG has joined #openstack-cinder | 21:51 | |
| thingee | xyang1: is this happening for liberty? https://blueprints.launchpad.net/cinder/+spec/emc-isilon-iscsi-volume-driver | 21:55 |
| *** rushil has quit IRC | 21:55 | |
| openstackgerrit | Tom Swanson proposed openstack/cinder: Dell: Added support for update_migrated_volume https://review.openstack.org/181160 | 21:56 |
| xyang1 | thingee: That is an old one, we are not contributing | 21:56 |
| *** david-lyle has joined #openstack-cinder | 21:56 | |
| xyang1 | thingee: Not sure if there is a way to abandon a bp | 21:56 |
| thingee | xyang1: I took care of it | 21:57 |
| thingee | xyang1: thanks | 21:57 |
| xyang1 | thingee: Thanks | 21:57 |
| xyang1 | thingee: There is a new driver bp for scaleio. That is for liberty | 21:58 |
| xyang1 | thingee: Did you see it? | 21:58 |
| Swanson | How do you go about identifying something as a potential bug fix for a previous release (kilo, in this case)? | 21:58 |
| thingee | xyang1: I believe I already approved it | 21:59 |
| xyang1 | thingee: Great, thanks | 21:59 |
| *** chlong has quit IRC | 21:59 | |
| *** dannywilson has quit IRC | 22:00 | |
| *** madskier has joined #openstack-cinder | 22:00 | |
| *** dannywil_ has joined #openstack-cinder | 22:01 | |
| *** dannywil_ has quit IRC | 22:02 | |
| *** ebalduf has quit IRC | 22:03 | |
| *** dannywilson has joined #openstack-cinder | 22:04 | |
| *** madskier has quit IRC | 22:05 | |
| *** emagana has quit IRC | 22:07 | |
| *** eharney has quit IRC | 22:07 | |
| *** akerr_ has quit IRC | 22:10 | |
| *** jungleboyj has joined #openstack-cinder | 22:10 | |
| *** akerr has joined #openstack-cinder | 22:11 | |
| *** abhiram_moturi has quit IRC | 22:18 | |
| *** abhiram_moturi has joined #openstack-cinder | 22:18 | |
| *** abhiram_moturi has quit IRC | 22:22 | |
| hemna | is that the error message you are seeing ? | 22:23 |
| *** abhiram_moturi has joined #openstack-cinder | 22:23 | |
| *** hodos has quit IRC | 22:24 | |
| *** abhiram_moturi has quit IRC | 22:27 | |
| *** abhiram_moturi has joined #openstack-cinder | 22:27 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Fix xxx=\n pep8 errors in volume_utils.py https://review.openstack.org/181179 | 22:28 |
| openstackgerrit | John Griffith proposed openstack/cinder: Check volume_backend in retype https://review.openstack.org/181079 | 22:28 |
| *** primechuck has quit IRC | 22:30 | |
| openstackgerrit | Vilobh Meshram proposed openstack/cinder: Bad link in API version details response https://review.openstack.org/180691 | 22:31 |
| *** vmtrooper has quit IRC | 22:31 | |
| *** abhiram_moturi has quit IRC | 22:34 | |
| *** ociuhandu has joined #openstack-cinder | 22:34 | |
| *** abhiram_moturi has joined #openstack-cinder | 22:34 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Check volume_backend in retype https://review.openstack.org/181079 | 22:36 |
| *** changbl has quit IRC | 22:36 | |
| *** david-lyle has quit IRC | 22:38 | |
| *** IanGovett has joined #openstack-cinder | 22:39 | |
| *** esker has quit IRC | 22:42 | |
| *** dannywilson has quit IRC | 22:44 | |
| *** dannywilson has joined #openstack-cinder | 22:46 | |
| *** akerr has quit IRC | 22:49 | |
| *** abhiram_moturi has quit IRC | 22:52 | |
| *** abhiram_moturi has joined #openstack-cinder | 22:53 | |
| *** abhiram_moturi has quit IRC | 22:55 | |
| *** abhiram_moturi has joined #openstack-cinder | 22:55 | |
| *** abhiram_moturi has quit IRC | 22:57 | |
| *** abhiram_moturi has joined #openstack-cinder | 22:58 | |
| *** abhiram_moturi has quit IRC | 23:02 | |
| *** AlkaD has quit IRC | 23:02 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:02 | |
| *** abhiram_moturi has quit IRC | 23:08 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:08 | |
| *** abhiram_moturi has quit IRC | 23:10 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:11 | |
| *** abhiram_moturi has quit IRC | 23:15 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:15 | |
| *** hemna is now known as hemnafk | 23:17 | |
| *** MentalRay has quit IRC | 23:18 | |
| *** abhiram_moturi has quit IRC | 23:24 | |
| *** chlong has joined #openstack-cinder | 23:24 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:24 | |
| *** bswartz has joined #openstack-cinder | 23:25 | |
| *** markvoelker has quit IRC | 23:26 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/cinder: Updated from global requirements https://review.openstack.org/178410 | 23:30 |
| *** abhiram_moturi has quit IRC | 23:35 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:35 | |
| openstackgerrit | John Griffith proposed openstack/cinder: Check volume_backend in retype https://review.openstack.org/181079 | 23:37 |
| *** abhiram_moturi has quit IRC | 23:39 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:40 | |
| *** abhiram_moturi has quit IRC | 23:44 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:44 | |
| *** abhiram_moturi has quit IRC | 23:46 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:46 | |
| *** abhiram_moturi has quit IRC | 23:48 | |
| *** abhiram_moturi has joined #openstack-cinder | 23:49 | |
| *** dims_ has joined #openstack-cinder | 23:50 | |
| *** dims has quit IRC | 23:52 | |
| *** mriedem is now known as mriedem_away | 23:54 | |
| *** takedakn has joined #openstack-cinder | 23:55 | |
| *** markvoelker has joined #openstack-cinder | 23:56 | |
| *** dustins has quit IRC | 23:56 | |
| *** markvoelker has quit IRC | 23:57 | |
| *** markvoelker has joined #openstack-cinder | 23:57 | |
| openstackgerrit | Vilobh Meshram proposed openstack/cinder: Bad link in API version details response https://review.openstack.org/180691 | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!
