| *** chlong has quit IRC | 00:03 | |
| *** yangyape_ has quit IRC | 00:17 | |
| *** yangyapeng has joined #openstack-cinder | 00:20 | |
| *** yangyapeng has quit IRC | 00:25 | |
| *** Julien-zte has joined #openstack-cinder | 00:36 | |
| *** wanghao_ has joined #openstack-cinder | 00:36 | |
| *** dave-mccowan has joined #openstack-cinder | 00:38 | |
| *** wanghao has quit IRC | 00:39 | |
| *** MVenesio has quit IRC | 00:47 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Cinder volume revert to snapshot https://review.openstack.org/391736 | 01:00 |
|---|---|---|
| *** fredli__ has joined #openstack-cinder | 01:01 | |
| *** dillaman has quit IRC | 01:08 | |
| *** jdillaman has quit IRC | 01:08 | |
| *** zhangguoqing has joined #openstack-cinder | 01:13 | |
| *** edisonxiang_ is now known as edisonxiang | 01:18 | |
| *** yangyapeng has joined #openstack-cinder | 01:21 | |
| *** jdillaman has joined #openstack-cinder | 01:21 | |
| *** dillaman has joined #openstack-cinder | 01:21 | |
| *** yangyape_ has joined #openstack-cinder | 01:25 | |
| openstackgerrit | Shunei Shiono proposed openstack/cinder master: NEC driver: Optimal path for non-disruptive backup https://review.openstack.org/474852 | 01:27 |
| *** yangyapeng has quit IRC | 01:28 | |
| *** kaisers_ has quit IRC | 01:33 | |
| *** Julien-zte has quit IRC | 01:36 | |
| *** KeithMnemonic1 has quit IRC | 01:39 | |
| *** lucasxu has joined #openstack-cinder | 01:39 | |
| *** lucasxu has quit IRC | 01:39 | |
| *** markvoelker has quit IRC | 01:40 | |
| *** markvoelker has joined #openstack-cinder | 01:41 | |
| *** markvoelker has quit IRC | 01:45 | |
| *** lhx has joined #openstack-cinder | 01:46 | |
| *** lhx is now known as Guest22742 | 01:47 | |
| *** gkadam__ has quit IRC | 01:47 | |
| openstackgerrit | jiamin proposed openstack/cinder master: DS8K driver: specify pool and lss in extra-specs https://review.openstack.org/435857 | 01:57 |
| *** bkopilov has quit IRC | 01:59 | |
| *** fredli__ has quit IRC | 02:02 | |
| *** adisky__ has quit IRC | 02:03 | |
| *** Guest22742 is now known as lhx__ | 02:05 | |
| *** dhinesh has quit IRC | 02:07 | |
| *** zhugaoxiao has joined #openstack-cinder | 02:13 | |
| *** mriedem has quit IRC | 02:24 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Cinder volume revert to snapshot https://review.openstack.org/391736 | 02:37 |
| *** mikal_ is now known as mikal | 03:15 | |
| *** bkopilov has joined #openstack-cinder | 03:17 | |
| *** zhangguoqing has quit IRC | 03:25 | |
| *** zhangguoqing has joined #openstack-cinder | 03:28 | |
| *** markvoelker has joined #openstack-cinder | 03:29 | |
| *** zhangguoqing has quit IRC | 03:36 | |
| *** Julien-zte has joined #openstack-cinder | 03:43 | |
| *** nicolasbock has joined #openstack-cinder | 03:46 | |
| *** gkadam__ has joined #openstack-cinder | 03:52 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Add ability to extend 'in-use' volume https://review.openstack.org/454287 | 03:55 |
| *** kaisers_ has joined #openstack-cinder | 04:00 | |
| *** jwcroppe has joined #openstack-cinder | 04:01 | |
| *** psachin has joined #openstack-cinder | 04:02 | |
| *** chhavi has joined #openstack-cinder | 04:03 | |
| *** links has joined #openstack-cinder | 04:07 | |
| *** dave-mccowan has quit IRC | 04:09 | |
| *** salv-orlando has joined #openstack-cinder | 04:12 | |
| *** salv-orlando has quit IRC | 04:13 | |
| *** salv-orlando has joined #openstack-cinder | 04:22 | |
| *** salv-orlando has quit IRC | 04:27 | |
| *** arnewiebalck_ has joined #openstack-cinder | 04:31 | |
| *** adisky__ has joined #openstack-cinder | 04:40 | |
| *** gouthamr has quit IRC | 04:43 | |
| *** arnewiebalck_ has quit IRC | 04:50 | |
| *** jerrygb has joined #openstack-cinder | 04:56 | |
| *** gkadam__ has quit IRC | 05:01 | |
| *** jerrygb has quit IRC | 05:01 | |
| *** salv-orlando has joined #openstack-cinder | 05:02 | |
| *** gkadam has joined #openstack-cinder | 05:05 | |
| *** links has quit IRC | 05:07 | |
| openstackgerrit | Cheng Yong Zhao proposed openstack/cinder master: Add gmcv support in SVC driver https://review.openstack.org/461191 | 05:07 |
| *** links has joined #openstack-cinder | 05:23 | |
| *** ChubYann has quit IRC | 05:27 | |
| *** salv-orlando has quit IRC | 05:39 | |
| *** salv-orlando has joined #openstack-cinder | 05:39 | |
| *** pewp has quit IRC | 05:40 | |
| *** e0ne has joined #openstack-cinder | 05:43 | |
| *** salv-orlando has quit IRC | 05:44 | |
| *** pewp has joined #openstack-cinder | 05:44 | |
| *** Julien-zte has quit IRC | 05:44 | |
| *** e0ne has quit IRC | 05:47 | |
| *** Julien-zte has joined #openstack-cinder | 05:48 | |
| *** Julien-zte has quit IRC | 05:50 | |
| *** Julien-zte has joined #openstack-cinder | 05:52 | |
| *** Julien-zte has quit IRC | 05:54 | |
| *** diablo_rojo has quit IRC | 05:57 | |
| *** Julien-zte has joined #openstack-cinder | 05:57 | |
| *** Julien-zte has quit IRC | 05:57 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Add ability to extend 'in-use' volume https://review.openstack.org/454287 | 06:04 |
| *** bkopilov has quit IRC | 06:06 | |
| *** Poornima_K has joined #openstack-cinder | 06:07 | |
| *** wanghao_ has quit IRC | 06:07 | |
| *** Poornima_K has quit IRC | 06:19 | |
| *** Julien-zte has joined #openstack-cinder | 06:22 | |
| *** lkuchlan has joined #openstack-cinder | 06:23 | |
| *** pcaruana has joined #openstack-cinder | 06:31 | |
| *** isaacb has joined #openstack-cinder | 06:32 | |
| *** Shunli has joined #openstack-cinder | 06:32 | |
| *** pgadiya has joined #openstack-cinder | 06:35 | |
| *** jerrygb has joined #openstack-cinder | 06:36 | |
| *** wanghao has joined #openstack-cinder | 06:36 | |
| *** bkopilov has joined #openstack-cinder | 06:42 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: [I18n] Don't convert exception.msg into string https://review.openstack.org/476022 | 06:51 |
| *** rcernin has joined #openstack-cinder | 06:59 | |
| *** markvoelker has quit IRC | 07:00 | |
| *** markvoelker has joined #openstack-cinder | 07:00 | |
| *** markvoelker has quit IRC | 07:01 | |
| *** markvoelker has joined #openstack-cinder | 07:01 | |
| *** markvoelker has quit IRC | 07:01 | |
| *** edmondsw has joined #openstack-cinder | 07:03 | |
| *** markvoelker has joined #openstack-cinder | 07:07 | |
| *** edmondsw has quit IRC | 07:08 | |
| *** tesseract has joined #openstack-cinder | 07:15 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: [i18n] Don't convert exception.msg into string https://review.openstack.org/476022 | 07:19 |
| openstackgerrit | Ashish Billore proposed openstack/cinder master: Include Usage for attachment-delete CLI https://review.openstack.org/476029 | 07:25 |
| *** links has quit IRC | 07:31 | |
| *** jerrygb has quit IRC | 07:34 | |
| *** zhangguoqing has joined #openstack-cinder | 07:35 | |
| openstackgerrit | Deepak Mourya proposed openstack/cinder-specs master: Remove support for py34. https://review.openstack.org/474445 | 07:35 |
| *** wanghao has quit IRC | 07:37 | |
| *** wanghao has joined #openstack-cinder | 07:39 | |
| *** links has joined #openstack-cinder | 07:44 | |
| *** arnewiebalck_ has joined #openstack-cinder | 07:44 | |
| openstackgerrit | TommyLike proposed openstack/python-cinderclient master: DNM. Test environment variables https://review.openstack.org/473165 | 07:45 |
| *** salv-orlando has joined #openstack-cinder | 07:54 | |
| *** zzzeek has quit IRC | 08:00 | |
| *** zzzeek has joined #openstack-cinder | 08:02 | |
| *** hjensas has quit IRC | 08:02 | |
| *** zhangguoqing has quit IRC | 08:09 | |
| *** arnewiebalck_ has quit IRC | 08:09 | |
| *** zhangguoqing has joined #openstack-cinder | 08:10 | |
| *** e0ne has joined #openstack-cinder | 08:10 | |
| *** sticker has quit IRC | 08:12 | |
| *** zhangguoqing has quit IRC | 08:13 | |
| *** e0ne has quit IRC | 08:13 | |
| openstackgerrit | Isaac Beckman proposed openstack/cinder master: IBM Storage - Fix vol create from snapshot in CG https://review.openstack.org/475793 | 08:14 |
| *** reveller has quit IRC | 08:15 | |
| openstackgerrit | jiamin proposed openstack/cinder master: DS8K: add replication consistency group support https://review.openstack.org/465877 | 08:17 |
| *** lpetrut has joined #openstack-cinder | 08:25 | |
| *** e0ne has joined #openstack-cinder | 08:27 | |
| *** chhavi has quit IRC | 08:28 | |
| *** e0ne has quit IRC | 08:33 | |
| *** zhangguoqing has joined #openstack-cinder | 08:39 | |
| *** gkadam_ has joined #openstack-cinder | 08:43 | |
| *** rcernin has quit IRC | 08:45 | |
| *** Poornima_K has joined #openstack-cinder | 08:46 | |
| *** gkadam has quit IRC | 08:46 | |
| *** rcernin has joined #openstack-cinder | 08:50 | |
| *** edmondsw has joined #openstack-cinder | 08:51 | |
| *** wanghao has quit IRC | 08:53 | |
| *** wanghao has joined #openstack-cinder | 08:54 | |
| *** edmondsw has quit IRC | 08:56 | |
| *** Julien-zte has quit IRC | 08:56 | |
| *** Julien-zte has joined #openstack-cinder | 08:58 | |
| *** hjensas has joined #openstack-cinder | 08:58 | |
| *** aspiers has joined #openstack-cinder | 09:00 | |
| *** salv-orlando has quit IRC | 09:02 | |
| *** salv-orlando has joined #openstack-cinder | 09:03 | |
| *** salv-orlando has quit IRC | 09:03 | |
| *** salv-orlando has joined #openstack-cinder | 09:06 | |
| *** salv-orlando has quit IRC | 09:08 | |
| *** salv-orlando has joined #openstack-cinder | 09:09 | |
| *** Poornima_K has quit IRC | 09:10 | |
| *** stakeda has quit IRC | 09:16 | |
| *** e0ne has joined #openstack-cinder | 09:18 | |
| *** chhavi has joined #openstack-cinder | 09:21 | |
| *** wanghao has quit IRC | 09:31 | |
| *** Shunli has quit IRC | 09:31 | |
| *** Poornima_K has joined #openstack-cinder | 09:41 | |
| *** e0ne has quit IRC | 09:50 | |
| *** e0ne has joined #openstack-cinder | 09:52 | |
| *** Julien-zte has quit IRC | 09:55 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - bootable volume errors https://review.openstack.org/472674 | 09:57 |
| *** gkadam_ has quit IRC | 10:06 | |
| *** gkadam_ has joined #openstack-cinder | 10:06 | |
| *** hjensas has quit IRC | 10:07 | |
| *** gkadam__ has joined #openstack-cinder | 10:07 | |
| openstackgerrit | ravi edpuganti proposed openstack/cinder master: Added CG capability to volume group in CoprHD https://review.openstack.org/459610 | 10:09 |
| *** gkadam_ has quit IRC | 10:11 | |
| *** salv-orl_ has joined #openstack-cinder | 10:12 | |
| *** salv-orlando has quit IRC | 10:14 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder master: WIP. Use 'ceph df' to calculate provisioned space https://review.openstack.org/476079 | 10:15 |
| Aurelgad1o | hello guys, I have an issue with cinder volume. with cinder-manage service list I can see my scheduler and my volume being online. When I create a new volume, I can see logs in the scheduler about my cinder-volume host being chosen for the volume, but I get nothing in the cinder-volume logs, the volume never gets created and stays in the "creating" status | 10:21 |
| *** gkadam__ is now known as gkadam | 10:21 | |
| Aurelgad1o | do you have any pointers on where i should look at ? | 10:21 |
| openstackgerrit | Isaac Beckman proposed openstack/cinder master: Ibm storage: Added replication group support https://review.openstack.org/468738 | 10:28 |
| *** yangyape_ has quit IRC | 10:30 | |
| *** Poornima_K has quit IRC | 10:33 | |
| *** edmondsw has joined #openstack-cinder | 10:39 | |
| *** zhonghua has quit IRC | 10:40 | |
| *** zhonghua has joined #openstack-cinder | 10:41 | |
| *** edmondsw has quit IRC | 10:45 | |
| *** chhavi has quit IRC | 10:47 | |
| *** yangyapeng has joined #openstack-cinder | 10:51 | |
| *** e0ne has quit IRC | 10:52 | |
| *** mvk has quit IRC | 10:54 | |
| *** bkopilov has quit IRC | 10:56 | |
| *** ericyoung has joined #openstack-cinder | 10:57 | |
| *** lhx__ has quit IRC | 10:57 | |
| *** zhangguoqing has quit IRC | 10:58 | |
| *** yangyapeng has quit IRC | 10:59 | |
| *** lhx__ has joined #openstack-cinder | 11:00 | |
| *** BlackDex_ is now known as BlackDex | 11:00 | |
| *** lhx__ has quit IRC | 11:10 | |
| *** lhx__ has joined #openstack-cinder | 11:10 | |
| openstackgerrit | Kushal Wathore proposed openstack/cinder master: HPE 3PAR: Adding driver capability https://review.openstack.org/474122 | 11:10 |
| openstackgerrit | Kushal Wathore proposed openstack/cinder master: HPE Lefthand: Fixed cloning operation https://review.openstack.org/462624 | 11:22 |
| *** mvk has joined #openstack-cinder | 11:23 | |
| openstackgerrit | Kushal Wathore proposed openstack/cinder master: 3PAR: Fixed image cache enabled capability. https://review.openstack.org/461744 | 11:25 |
| openstackgerrit | Merged openstack/python-cinderclient master: Fix cmd options for updating a quota class https://review.openstack.org/453134 | 11:28 |
| *** abishop has joined #openstack-cinder | 11:35 | |
| *** yangyapeng has joined #openstack-cinder | 11:36 | |
| *** yangyapeng has quit IRC | 11:41 | |
| *** yangyapeng has joined #openstack-cinder | 11:41 | |
| *** dave-mccowan has joined #openstack-cinder | 11:41 | |
| *** reveller has joined #openstack-cinder | 11:47 | |
| openstackgerrit | Merged openstack/cinder master: Fix NoneType has no attribute get error https://review.openstack.org/471672 | 11:49 |
| openstackgerrit | Merged openstack/cinder master: Replace deprecated BaseException.message with six.text_type https://review.openstack.org/471299 | 11:51 |
| *** isaacb_ has joined #openstack-cinder | 11:52 | |
| *** isaacb has quit IRC | 11:53 | |
| *** e0ne has joined #openstack-cinder | 11:56 | |
| *** zhangguoqing has joined #openstack-cinder | 12:00 | |
| *** zhangguoqing has quit IRC | 12:06 | |
| *** pgadiya has quit IRC | 12:14 | |
| danpawlik | jbernard, jungleboyj, sdatko, smcginnis: I added release notes https://review.openstack.org/#/c/456672/ . I hope it should be good right now | 12:19 |
| danpawlik | can you review :) ? | 12:19 |
| danpawlik | thanks! | 12:19 |
| *** edmondsw has joined #openstack-cinder | 12:21 | |
| *** isaacb_ has quit IRC | 12:25 | |
| *** isaacb_ has joined #openstack-cinder | 12:27 | |
| *** zhangguoqing has joined #openstack-cinder | 12:28 | |
| *** bkopilov has joined #openstack-cinder | 12:32 | |
| openstackgerrit | Eric Young proposed openstack/cinder master: ScaleIO Driver: Fix for using scheduler specified pool https://review.openstack.org/476116 | 12:39 |
| *** isaacb__ has joined #openstack-cinder | 12:40 | |
| *** isaacb__ has quit IRC | 12:40 | |
| *** isaacb_ has quit IRC | 12:42 | |
| jungleboyj | danpawlik: Reviewed. Thanks for dealing with all my harassment. :-) | 12:48 |
| danpawlik | jungleboyj: Its good that you look on the details. Its really important and a right way to have a good/clean code in repo :) | 12:51 |
| *** eharney has joined #openstack-cinder | 12:51 | |
| jungleboyj | danpawlik: You are right. Appreciate that you understand that! | 12:53 |
| *** salv-orl_ has quit IRC | 12:54 | |
| *** mriedem has joined #openstack-cinder | 12:56 | |
| *** mdovgal has quit IRC | 12:56 | |
| *** salv-orlando has joined #openstack-cinder | 12:56 | |
| *** liverpooler has joined #openstack-cinder | 12:57 | |
| *** salv-orlando has quit IRC | 12:57 | |
| *** pewp has quit IRC | 12:57 | |
| *** salv-orlando has joined #openstack-cinder | 12:58 | |
| *** dave-mccowan has quit IRC | 12:58 | |
| *** jerrygb has joined #openstack-cinder | 13:00 | |
| *** lhx__ has quit IRC | 13:03 | |
| *** jwcroppe has quit IRC | 13:03 | |
| *** jwcroppe has joined #openstack-cinder | 13:03 | |
| *** jwcroppe has quit IRC | 13:03 | |
| *** jwcroppe has joined #openstack-cinder | 13:04 | |
| *** pewp has joined #openstack-cinder | 13:04 | |
| *** jwcroppe has quit IRC | 13:06 | |
| *** gouthamr has joined #openstack-cinder | 13:07 | |
| *** lucasxu has joined #openstack-cinder | 13:07 | |
| *** salv-orlando has quit IRC | 13:08 | |
| *** salv-orlando has joined #openstack-cinder | 13:12 | |
| *** cknight has joined #openstack-cinder | 13:14 | |
| openstackgerrit | Ivan Kolodyazhny proposed openstack/cinder master: WIP. Do not fail if RBD image is broken https://review.openstack.org/475400 | 13:17 |
| *** cknight has quit IRC | 13:18 | |
| *** jwcroppe has joined #openstack-cinder | 13:30 | |
| *** eric_young has joined #openstack-cinder | 13:31 | |
| openstackgerrit | Jeremy Zhang proposed openstack/cinder master: Fix assertTrue in functional tests https://review.openstack.org/475661 | 13:32 |
| *** ericyoung has quit IRC | 13:33 | |
| *** lhx has joined #openstack-cinder | 13:38 | |
| *** lhx is now known as Guest79242 | 13:38 | |
| *** xyang1 has joined #openstack-cinder | 13:38 | |
| openstackgerrit | Jeremy Zhang proposed openstack/cinder master: Fix assertTrue in functional tests https://review.openstack.org/475661 | 13:39 |
| openstackgerrit | TommyLike proposed openstack/python-cinderclient master: Enabled like filter support in client https://review.openstack.org/466214 | 13:40 |
| *** dave-mccowan has joined #openstack-cinder | 13:41 | |
| *** reveller has quit IRC | 13:43 | |
| *** cknight has joined #openstack-cinder | 13:44 | |
| *** catintheroof has joined #openstack-cinder | 13:45 | |
| *** reveller has joined #openstack-cinder | 13:46 | |
| tommylikehu | anyone who is familiar with i18n could take a look at this patch? https://review.openstack.org/#/c/474547/ thanks in advance :) | 13:47 |
| *** diablo_rojo has joined #openstack-cinder | 13:47 | |
| *** zhangguoqing has quit IRC | 13:48 | |
| *** salv-orlando has quit IRC | 13:49 | |
| *** zhangguoqing has joined #openstack-cinder | 13:51 | |
| *** lhx_ has joined #openstack-cinder | 13:54 | |
| *** Guest79242 has quit IRC | 13:57 | |
| *** adisky__ has quit IRC | 13:59 | |
| *** jerrygb has quit IRC | 14:04 | |
| *** crushil has joined #openstack-cinder | 14:04 | |
| *** chlong has joined #openstack-cinder | 14:07 | |
| *** gkadam has quit IRC | 14:08 | |
| *** felipemonteiro has joined #openstack-cinder | 14:14 | |
| *** felipemonteiro_ has joined #openstack-cinder | 14:15 | |
| *** Julien-zte has joined #openstack-cinder | 14:17 | |
| *** felipemonteiro has quit IRC | 14:19 | |
| tommylikehu | ping geguileo | 14:22 |
| tommylikehu | what's the issue we have in the wsgi file at present? https://github.com/openstack/cinder/blob/master/cinder/api/openstack/wsgi.py | 14:23 |
| *** dustins has joined #openstack-cinder | 14:30 | |
| *** Julien-zte has quit IRC | 14:32 | |
| *** Julien-z_ has joined #openstack-cinder | 14:34 | |
| *** reveller has quit IRC | 14:36 | |
| *** Julien-z_ has quit IRC | 14:37 | |
| *** gouthamr has quit IRC | 14:37 | |
| *** gouthamr has joined #openstack-cinder | 14:37 | |
| e0ne | jbernard, eharney: hi guys, are you around? | 14:39 |
| jbernard | e0ne: nope | 14:39 |
| e0ne | :( | 14:39 |
| jbernard | e0ne: jk, im here! XD | 14:39 |
| e0ne | I would like to discuss https://bugs.launchpad.net/cinder/+bug/1698786 | 14:39 |
| openstack | Launchpad bug 1698786 in Cinder "cinder-volume fails on start when rbd pool contains partially deleted images" [Undecided,In progress] - Assigned to Ivan Kolodyazhny (e0ne) | 14:39 |
| * jbernard 's still got it | 14:39 | |
| jbernard | e0ne: sure | 14:40 |
| e0ne | there are two proposed fixes: https://review.openstack.org/475400 and https://review.openstack.org/476079 | 14:40 |
| *** markvoelker has quit IRC | 14:40 | |
| e0ne | the first one looks like workaround | 14:40 |
| e0ne | but the second one changes how we calculate capacity now | 14:40 |
| jbernard | gimme a sec, looking | 14:41 |
| e0ne | sure | 14:41 |
| *** reveller has joined #openstack-cinder | 14:42 | |
| jbernard | i cannot remember why we iterate on volumes… that would have been a nice place for a comment | 14:43 |
| jbernard | e0ne: the results are identical? | 14:43 |
| *** Julien-zte has joined #openstack-cinder | 14:43 | |
| e0ne | jbernard, nope. 'ceph df' shows real provisioned size | 14:44 |
| e0ne | jbernard: the current method returns allocated space | 14:44 |
| jbernard | ahh right, we want allocated, no? | 14:44 |
| e0ne | we want, if I didn't miss anything | 14:45 |
| jbernard | else we loose our ability to overprovision a thin backend | 14:45 |
| e0ne | let me find review request with current implementation. there were some useful comments | 14:45 |
| jbernard | or maybe we handle that elsewhere | 14:45 |
| e0ne | here is it https://review.openstack.org/#/c/178262/ | 14:46 |
| e0ne | jbernard: here is your comment https://review.openstack.org/#/c/178262/1/cinder/volume/drivers/rbd.py | 14:47 |
| eharney | i don't think we want to shell out to "ceph df" | 14:47 |
| e0ne | eharney: I agree with you, but I didn't find python bindings for it | 14:48 |
| jbernard | Rados.get_cluster_stats() | 14:48 |
| jbernard | http://docs.ceph.com/docs/master/rados/api/python/ | 14:48 |
| jbernard | that may not offer stats per-pool though :/ | 14:49 |
| geguileo | tommylikehu: I realized during lunch that I was mistaken, was about to mention it in the review | 14:50 |
| geguileo | lol | 14:50 |
| jbernard | but there are two issues, 1) volumes shouldn't be able to be left in a partially deleted state | 14:50 |
| jbernard | and 2) what is _get_usage_info() expected to return? | 14:50 |
| e0ne | jbernard: you're right. Rados.get_cluster_stats doesn't return per pool stats | 14:51 |
| eharney | _get_usage_info is populating the normal get_volume_stats info | 14:51 |
| e0ne | jbernard: they shouldn't. but in real world something goes wrong:( | 14:53 |
| e0ne | jbernard: probably, we need some periodic job to clean up such volumes | 14:53 |
| jbernard | e0ne: could we at least detect them and move them to lost+found (or something to that effect) | 14:54 |
| e0ne | jbernard: I'm working on such patch | 14:54 |
| e0ne | jbernard: but it's a bit complicated because we have to figure out what volumes are failed to delete and what are deleting now | 14:55 |
| *** markvoelker has joined #openstack-cinder | 14:57 | |
| *** rcernin has quit IRC | 15:07 | |
| *** jwcroppe has quit IRC | 15:07 | |
| e0ne | jbernard, eharney: I'll be happy to get any feedback from you to get this issue fixed | 15:09 |
| jbernard | e0ne: personally i think skipping images that raise an exception is okay | 15:11 |
| jbernard | e0ne: unless we're not supposed to report allocated space | 15:11 |
| jbernard | e0ne: but i think we are | 15:11 |
| jbernard | a librados call for pool stats would be nice | 15:11 |
| *** zhangguoqing has quit IRC | 15:12 | |
| e0ne | jbernard: do we have such api for librados? | 15:12 |
| *** pcaruana has quit IRC | 15:12 | |
| jbernard | e0ne: not that i know of, at least by the docs, i need to look closer | 15:13 |
| *** psachin has quit IRC | 15:13 | |
| openstackgerrit | Merged openstack/cinder master: Fix assertTrue in functional tests https://review.openstack.org/475661 | 15:13 |
| *** chhavi has joined #openstack-cinder | 15:18 | |
| *** dillaman has quit IRC | 15:19 | |
| *** dillaman has joined #openstack-cinder | 15:21 | |
| Aurelgad1o | also iirc ceph df shows the raw storage, it does not take into account the replica overhead | 15:28 |
| Aurelgad1o | I beg your pardon for reposting but now that north america is awake ... | 15:29 |
| e0ne | Aurelgad1o: how big is overhead for replica? | 15:29 |
| Aurelgad1o | I have an issue with cinder volume. with cinder-manage service list I can see my scheduler and my volume being online. When I create a new volume, I can see logs in the scheduler about my cinder-volume host being chosen for the volume, but I get nothing in the cinder-volume logs, the volume never gets created and stays in the "creating status" | 15:29 |
| Aurelgad1o | e0ne: if you have three OSDs of 1 TB, you'll see free space as 3TB. By default, ceph replicates data 3 times, but it depends on the pool's configuration | 15:30 |
| *** sudipto has joined #openstack-cinder | 15:31 | |
| *** sudipto_ has joined #openstack-cinder | 15:31 | |
| Aurelgad1o | a ceph admin can also set quotas on pools : http://docs.ceph.com/docs/kraken/rados/operations/pools/#set-pool-quotas | 15:32 |
| e0ne | Aurelgad1o: it could be an issue with messaging (oslo.messaging, rabbitmq) in your case | 15:32 |
| Aurelgad1o | e0ne: I don't see messages staying in the queues, but how can I be sure messages are posted in the queues ? the DEBUG verbosity does not seems to be enough for me to see that | 15:32 |
| e0ne | Aurelgad1o: do you have any errors in scheduler logs? | 15:33 |
| Aurelgad1o | e0ne: nope, I see the volumebeing scheduled to the host I want to and there is no error | 15:34 |
| *** lpetrut has quit IRC | 15:36 | |
| *** felipemonteiro_ has quit IRC | 15:38 | |
| *** mchlumsky has joined #openstack-cinder | 15:39 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - Replication, replacing SMI-S with REST https://review.openstack.org/466027 | 15:41 |
| *** danpawlik is now known as danpawlik_ | 15:43 | |
| *** crushil has quit IRC | 15:44 | |
| mchlumsky | Hello. I have a question regarding the volume_clear* options in cinder.conf. Is the volume_clear functionality implemented in each driver? | 15:46 |
| e0ne | mchlumsky: hi. no, it's implemented only fror lvm and unsupported bdd drivers | 15:51 |
| *** wewe0901 has joined #openstack-cinder | 15:51 | |
| e0ne | Aurelgad1o: :(. looks like something with messaging, but I'm not sure | 15:51 |
| *** crushil has joined #openstack-cinder | 15:52 | |
| mchlumsky | e0ne, thanks! | 15:53 |
| *** salv-orlando has joined #openstack-cinder | 15:55 | |
| *** salv-orlando has quit IRC | 15:56 | |
| *** salv-orlando has joined #openstack-cinder | 15:57 | |
| *** markvoelker has quit IRC | 15:58 | |
| *** markvoelker has joined #openstack-cinder | 15:59 | |
| *** diablo_rojo has quit IRC | 16:00 | |
| geguileo | jungleboyj: I broke disconnect for StorwizeSVCISCSIDriver and some other drivers with latest os-brick changes | 16:00 |
| geguileo | jungleboyj: Any change you could test the fix? https://review.openstack.org/#/c/475760/ | 16:00 |
| *** Apoorva has joined #openstack-cinder | 16:00 | |
| *** diablo_rojo has joined #openstack-cinder | 16:01 | |
| jungleboyj | I don't have access to that hardware anymore. hemna ^ | 16:01 |
| *** gnarld_ is now known as cFouts | 16:01 | |
| *** tesseract has quit IRC | 16:03 | |
| *** aarefiev is now known as aarefiev_afk | 16:07 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - QoS, replacing SMI-S with REST https://review.openstack.org/456759 | 16:08 |
| *** felipemonteiro has joined #openstack-cinder | 16:09 | |
| *** felipemonteiro_ has joined #openstack-cinder | 16:10 | |
| *** salv-orl_ has joined #openstack-cinder | 16:12 | |
| *** lkuchlan has quit IRC | 16:12 | |
| *** felipemonteiro has quit IRC | 16:14 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - Compression, replacing SMI-S with REST https://review.openstack.org/460191 | 16:14 |
| *** salv-orlando has quit IRC | 16:15 | |
| *** Julien-zte has quit IRC | 16:21 | |
| *** Julien-zte has joined #openstack-cinder | 16:24 | |
| *** Julien-zte has quit IRC | 16:24 | |
| *** jwcroppe has joined #openstack-cinder | 16:35 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - Replication, replacing SMI-S with REST https://review.openstack.org/466027 | 16:40 |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - Live Migration, replacing SMI-S with REST https://review.openstack.org/472279 | 16:43 |
| *** jwcroppe has quit IRC | 16:44 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Fix i18n issues https://review.openstack.org/474547 | 16:44 |
| *** mvk has quit IRC | 16:56 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - implement Generic Volume Group feature https://review.openstack.org/474546 | 16:59 |
| *** Apoorva_ has joined #openstack-cinder | 17:02 | |
| tbarron | winston-d: was just wondering what the multinode SIGHUP issue is exactly | 17:03 |
| jungleboyj | People talk to much! | 17:04 |
| eharney | i'm still lost on why there's a concern there about multinode and HA too | 17:04 |
| tbarron | winston-d: clearly one has to change config on multiple nodes and deliver a SIGHUP on each | 17:04 |
| jungleboyj | @!t | 17:04 |
| pewp | jungleboyj (╯°□°)╯︵ ┻━┻ | 17:04 |
| tbarron | winston-d: and if you do that there will be windows when some nodes are running old config and some running nes config | 17:04 |
| tbarron | winston-d: do we have options that cause a problem in those windows? | 17:05 |
| jungleboyj | tbarron: Right, so you are saying the only option is to shut everone down and then restart? | 17:05 |
| tbarron | jungleboyj: I'm not *saying* that, but I'm asking if that's true and why | 17:05 |
| *** Apoorva has quit IRC | 17:05 | |
| *** lhx_ has quit IRC | 17:06 | |
| tbarron | jungleboyj: there are clearly windows for inconsistent options, I'm just wondering what they are and do they cause trouble | 17:06 |
| winston-d | having to restart cinder to reload new cfg value, means you can trade a small down time with consistency. | 17:06 |
| tbarron | jungleboyj: winston-d for example an option to change logging will not cause a problem in those windows | 17:06 |
| *** harlowja has joined #openstack-cinder | 17:07 | |
| winston-d | node 1/node 2 are HA pair, you change cfg for both nodes, you shutdown 1 and 2 and then start them one by one. | 17:07 |
| tbarron | jungleboyj: winston-d but if you were running active-active volume service and changed an option that affects actual backend behavior ... | 17:07 |
| winston-d | no inconsistent window | 17:07 |
| tbarron | winston-d: "jinx" | 17:07 |
| tbarron | winston-d: do you know of other problematic cases than active/active volume service? | 17:08 |
| tbarron | winston-d: problems for multiple api, scheduler, or backup services? | 17:09 |
| mchlumsky | @eric_young, Hi! It looks like ScaleIO doesn't wipe the data after we delete a volume in OpenStack (unless it's something we need to configure in ScaleIO itself). We've seen newly created volumes with pre-existing filesystems on them. Can the volume_clear feature be implemented in the ScaleIO driver? | 17:09 |
| tbarron | winston-d: or problems across services (seems unlikely b/c it's loosely coupled rpc stuff) | 17:09 |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - bootable volume errors https://review.openstack.org/472674 | 17:10 |
| *** lbragstad has joined #openstack-cinder | 17:10 | |
| winston-d | tbarron: I don't, I'm trying to see if I can think of any. | 17:11 |
| *** e0ne has quit IRC | 17:11 | |
| tommylikehu | hey geguileo , does this what are you prefer? https://review.openstack.org/#/c/474547/ | 17:11 |
| pewp | hemna ( ´ ▽ ` )ﾉ | 17:11 |
| *** chhavi has quit IRC | 17:12 | |
| lbragstad | diablo_rojo: fwiw - one thing that's helped keystone cut down on cruft in our configuration file is the use of resource-option | 17:12 |
| openstackgerrit | Merged openstack/cinder master: VMAX driver - Retype, replacing SMI-S with REST https://review.openstack.org/456757 | 17:12 |
| winston-d | if inconsistent API leads to potentially inconsistent data plane change, it'd be considered as bad as inconsistent A/A c-vols | 17:13 |
| lbragstad | diablo_rojo: we noticed it when we were ending up with whitelisted user-ids in our config file for various things | 17:13 |
| lbragstad | diablo_rojo: so - we implemented https://github.com/openstack/keystone/commit/1896d1ba0d24d3780ce8c7652fa4c4378a02255d | 17:14 |
| lbragstad | diablo_rojo: which allowed us to deprecate and remove some of our configuration cruft - https://github.com/openstack/keystone/commit/9844fa1e264fe52276d9bd804b26b518ef3aaa24 | 17:14 |
| tbarron | winston-d: so I have a theoretical feel for how there *could* be problems in this regard (maybe) but | 17:15 |
| tbarron | winston-d: do you actually have any concrete known options that cause such problems? | 17:15 |
| tbarron | winston-d: even with active-active volume service? | 17:15 |
| *** links has quit IRC | 17:15 | |
| tbarron | winston-d: for the sake of argument we can assume active-active volume service for the same backend | 17:16 |
| tbarron | winston-d: and SIGHUP one volume service, then wait ten minutes and SIGHUP the other | 17:16 |
| tbarron | in both cases changing from config A to config A' | 17:16 |
| tbarron | winston-d: what options make this break things in that 10 minute window? | 17:16 |
| diablo_rojo | lbragstad, ooOOoo good to know | 17:18 |
| lbragstad | diablo_rojo: in case that helps at all | 17:18 |
| openstackgerrit | Merged openstack/cinder master: Cinder volume revert to snapshot https://review.openstack.org/391736 | 17:19 |
| winston-d | tbarron: i probably need more than 10 mins to think of one. ;) | 17:19 |
| tbarron | winston-d: fair enough | 17:20 |
| tbarron | winston-d: but if it's just a *theoretical* possibility that there *could* be such options, then maybe | 17:20 |
| tbarron | winston-d: it would be sufficient to make sure that new options don't have that kind of problem in code review | 17:21 |
| *** jwcroppe has joined #openstack-cinder | 17:21 | |
| tbarron | winston-d: the reason I say that is that tightly coupling service instances on multiple nodes in that way seems to me a poor idea in the first place | 17:21 |
| *** _alastor1 is now known as _alastor_ | 17:21 | |
| hemna | changing the default volume type will lead to inconsistent behavior. | 17:22 |
| eharney | the other question there is, if there are options that would break like that, do they break already if people start/stop/restart services in the right order? | 17:22 |
| *** mvk has joined #openstack-cinder | 17:22 | |
| eharney | inconsistent behavior like a different default volume type seems like a reasonable thing to have happen in this circumstance, people need to know that if they change config, all of the services need to reload it | 17:23 |
| eharney | that isn't any different from what happens with service configs today | 17:23 |
| hemna | there are many things in the cinder.conf changes that would break things | 17:24 |
| hemna | db url, transport_url | 17:24 |
| hemna | stuffs like that | 17:25 |
| tbarron | hemna: that would break things on multi-node today if we don't stop, edit, start in the right order? | 17:25 |
| hemna | yah | 17:25 |
| hemna | if you are dumb enough to change that | 17:25 |
| hemna | and not restart all nodes | 17:25 |
| hemna | do the cinder.conf quota settings do anything after the first start? | 17:25 |
| tbarron | hemna: do these give more than transitory errors if we stop and restart quickly? | 17:26 |
| winston-d | well, human error is always there, i don't think it's fair to go by flawed design because human err is inevitable. | 17:27 |
| hemna | what concerns me about this is that we are all not sure what will happen at this point. | 17:27 |
| hemna | and no way to test it | 17:27 |
| tbarron | hemna: so is using SIGHUP (delivering a storm of them in random order) to all the nodes running cinder services more dangerous than delivering a similarly random 'systemctl stop <service'; sed-cmd; 'systemctl start service' via ssh to all the relevant nodes? | 17:28 |
| jgriffith | so not to interrupt here... | 17:28 |
| jgriffith | but I'm going to anyway | 17:28 |
| jgriffith | :) | 17:28 |
| tbarron | my point is that it's not the SIGHUP that is the issue | 17:28 |
| * tbarron pauses a sec | 17:29 | |
| jgriffith | in a standard multi-node today there's not really a problem here because there's no shared state | 17:29 |
| jgriffith | the example of the db node or rpc node is kinda silly | 17:29 |
| tbarron | pewp | 17:29 |
| hemna | yah it is, but it's still valid | 17:29 |
| jgriffith | but in an A/A c-vol scenario there are a number of pit-falls that are valid | 17:30 |
| *** sudipto_ has quit IRC | 17:30 | |
| *** sudipto has quit IRC | 17:30 | |
| jgriffith | your driver configs need to match | 17:30 |
| jgriffith | and the example of default volume-type is an interesting point IMO | 17:30 |
| tbarron | i'm looking for a concrete example of harm by mismatch | 17:31 |
| tbarron | not saying it's not there, just think it would be useful | 17:31 |
| jgriffith | default volume type is a pretty good one | 17:31 |
| jgriffith | IMO | 17:31 |
| jgriffith | misconfigured device option | 17:31 |
| jgriffith | some devices not a big deal... but say Netapp for example that has 22 options or something like that | 17:31 |
| tbarron | what's exactly the harm if that is different for a brief period? | 17:31 |
| winston-d | jgriffith: if someone changed the 'sf_vol_prefix' on one node for SF driver, and the other node only got updated with new value after 10 mins, would that be a problem? | 17:31 |
| jgriffith | there's potential there for bad things | 17:32 |
| jgriffith | winston-d yes | 17:32 |
| jgriffith | winston-d that would mean that a/a wouldn't work | 17:32 |
| winston-d | jgriffith: node1, 'uuid' as prefix, a vol created with that, node2, 'cinder' as prefix. | 17:32 |
| jgriffith | because you'd be looking for the wrong volume-name | 17:32 |
| tbarron | so you would have some "old" prefixes and some "new" prefixes anyways unless SF converts all the old to new | 17:32 |
| jgriffith | you're screwed | 17:32 |
| smcginnis | diablo_rojo, jgriffith: FYI, Doug is trying to get pieces in place to be able to generate config files. This is part of that effort: https://review.openstack.org/#/c/474307/ | 17:33 |
| jgriffith | tbarron it doesn't work that way sadly (or happily actually) | 17:33 |
| tbarron | playing devil's advocate, your driver should convert old to new | 17:33 |
| winston-d | user vol A got created by node1, and then he wanted to create a snapshot/or delete the vol, but that request got handled by node2. | 17:33 |
| jgriffith | tbarron ummm... no; actually it shouldn't | 17:33 |
| eharney | the sf_vol_prefix case is one where you can't change it for a deployment that has volumes even on a single node deployment | 17:33 |
| tbarron | jgriffith: i expect you are right | 17:33 |
| winston-d | vol A's the status would become 'error_deleting' if request was a deleete? | 17:33 |
| jgriffith | eharney you're correct | 17:33 |
| jgriffith | at least kinda | 17:34 |
| eharney | there are a handful of options in that category, i think | 17:34 |
| diablo_rojo | smcginnis, thanks for the heads up | 17:34 |
| * diablo_rojo opens the link | 17:34 | |
| jgriffith | eharney you're right.. but I do think there is at least some things in an A/A setup that should be considered a bit | 17:34 |
| jgriffith | and what I do think would be interesting is a 'shared-config' | 17:35 |
| jgriffith | so you introduce a config option that says "a/a pair with xyz" or whatever | 17:35 |
| tbarron | so volume prefix isn't safe, but it's never safe to change anyways, right? | 17:35 |
| jgriffith | and the config is shared | 17:35 |
| tbarron | so it's not the concrete example why SIGHUP is bad | 17:35 |
| jgriffith | tbarron well, it's somewhat different... it's not necessarily safe in any scenario unless you're doing it knowingly; but in A/A it has all sorts of weird effects | 17:36 |
| jgriffith | depending on which c-vol service receives the request determines what the behavior/results will be | 17:36 |
| jgriffith | as opposed to a traditional deployment it's just consistent and relatively easy to trouble-shoot and figure out | 17:37 |
| jgriffith | I'm not arguing one side or the other here, just presenting some thoughts | 17:38 |
| tbarron | yeah, with A/A if a backend can't handle "old" config after it has seen "new" config then one should shut them all down, change config for all, and start | 17:38 |
| jgriffith | what about the shared-config idea? | 17:38 |
| jgriffith | anybody have thoughts on that? | 17:38 |
| tbarron | jgriffith: explain it a bit more | 17:38 |
| jgriffith | well, if I'm saying this is an A/A c-vol setup; then I most likely want to use the same config on both nodes right? | 17:39 |
| tbarron | in manila I have some private config that I want to share between multiple ganesha servers on different nodes BTW | 17:39 |
| tbarron | will probably put it (json) in a rados object | 17:39 |
| jgriffith | there's nothing to prevent sharing a config across nodes via rpc I don't think | 17:40 |
| jgriffith | or... | 17:40 |
| jgriffith | adding the ability to perform a check/comparison | 17:40 |
| jgriffith | so as part of init you do a compare of the two and make sure they match | 17:40 |
| jgriffith | if they don't, you can log the info/deltas or maybe error out | 17:40 |
| jgriffith | error out and fail to init may not be the way to go at first | 17:41 |
| jgriffith | but logging error info stating "Hey! These configs for the driver are different, and here's the things that don't match!!! You should look at these and make sure this is ok" | 17:41 |
| jgriffith | It's just a usability thing | 17:42 |
| winston-d | well, the problem is not just config failed to be deployed to pair of nodes simultaneously. it's more about if pair of nodes can load new config into memory and start acting with new behavior. | 17:42 |
| winston-d | with as minimum as possible drift/inconsistent window. | 17:42 |
| jgriffith | winston-d you mean like the config on one node was flat out wrong/incopatible? | 17:42 |
| jgriffith | incompatable | 17:43 |
| jgriffith | with the backend device? | 17:43 |
| jgriffith | so the drift window... IMO that circles back to the only solution being a shared config | 17:43 |
| jgriffith | I don't know of another way to eliminate that window completely | 17:44 |
| *** jwcroppe has quit IRC | 17:45 | |
| winston-d | take sf_vol_prefix for example, 'uuid' is changed to 'cinder', the cinder.conf on both nod1 and node2 has been changed successfully at almost the same time, but node1 is doing somethign heavy with high load, it took it 30 secs before the new value is applied while node2 takes the new value alsmost instantly. | 17:45 |
| winston-d | my point is exactly that there is no way to eliminate that window. | 17:46 |
| jgriffith | winston-d yeah... there really isn't | 17:46 |
| jgriffith | even my idea would still have a small window | 17:46 |
| jgriffith | anything already on the queue it's just too bad | 17:47 |
| winston-d | is 1 sec window small enough? in our peak hour, we might have 20~40 vol being created in 1 sec. | 17:47 |
| jgriffith | probably not | 17:47 |
| eric_young | mchlkumsky: That sounds odd to me, can you follow up with your Dell EMC Support channels. I would imagine that creating a new volume in SIO would NOT give you back an existing volume with a filesystem | 17:48 |
| winston-d | the behavior of dynamic config reload is not predictable becuase of load or other factors. | 17:49 |
| patrickeast | it might just be something on a case-by-case basis for the config options... they aren't all equal, some you really might need to stop all requests or do some more complex workflow | 17:49 |
| jgriffith | patrickeast +1 | 17:49 |
| patrickeast | imo adding a reload without restart is just a tool for advanced users... but a really dangerous one | 17:49 |
| tbarron | patrickeast: yeah | 17:49 |
| winston-d | the safest way is to trade down time for consistency. | 17:49 |
| jgriffith | I guess the only option there would be flush the queue, wait, restart | 17:49 |
| winston-d | a maintenance windows and notification email save us all. | 17:50 |
| jgriffith | winston-d seems reasonable to me | 17:50 |
| tbarron | but it would be real handy for when you want to turn on debug | 17:50 |
| tbarron | and stuff like that | 17:50 |
| tbarron | maybe documenting when it's safe and when not makes sense | 17:51 |
| patrickeast | tbarron: +1 | 17:51 |
| patrickeast | things like that, or adjusting timeouts | 17:51 |
| patrickeast | it makes total sense | 17:51 |
| winston-d | tbarron: that i totally agree. you know how painful to restart a c-vol hosts a few thousands vols. | 17:51 |
| smcginnis | tbarron: https://review.openstack.org/#/c/445885/ | 17:51 |
| jgriffith | smcginnis you beat me to it! | 17:51 |
| smcginnis | Hah, and I'm not even really here. :P | 17:52 |
| winston-d | but in production, we want consistent and predictable behavior to minize the risk. | 17:52 |
| jgriffith | smcginnis either shows that you're "that good" or that I'm "that bad" | 17:52 |
| jgriffith | or both :) | 17:52 |
| smcginnis | Let's go with neither and that I'm just bad at focusing on training material. ;) | 17:52 |
| jgriffith | haha | 17:53 |
| jgriffith | shhhh.. they might be listening! | 17:53 |
| jgriffith | so what I've gathered from this long discussion that I've admittedly only partially followed and understood everything about is; I don't think dynamic config is such a hot idea | 17:54 |
| jgriffith | and by that I mean, updating config without a restart | 17:54 |
| smcginnis | jgriffith: I think this is actually more useful in that we need to explicitly decide each option that can be changed: https://review.openstack.org/#/c/464028/ | 17:55 |
| winston-d | jgriffith: that's my take as well. | 17:55 |
| jgriffith | smcginnis that's interesting... hadn't seen it | 17:55 |
| jgriffith | I was about to say that we could selectively choose things to "allow" and add it to geguileo 's patch for log level | 17:56 |
| jgriffith | but the mutable config thing is interesting | 17:56 |
| * jgriffith doesn't completely grok it yet | 17:56 | |
| smcginnis | jgriffith: Yep, this would be the first step, then we can mark specific options that we would want to be able to change. | 17:57 |
| jgriffith | smcginnis winston-d tbarron that seems like a reasonable compromise no? | 17:57 |
| tbarron | understanding which options are safe to change without stopping services on all nodes and updating all config before restarting would be worthwhile in itself | 17:57 |
| winston-d | mutable without restart? | 17:57 |
| jgriffith | winston-d yes, but only white-listed options | 17:57 |
| smcginnis | FWIW, I don't think this oslo.config option has been really communicated as a new option folks should look at. | 17:58 |
| tbarron | just not allowing SIGHUP doesn't by itself guarantee that ops will do the right thing anyways | 17:58 |
| jgriffith | tbarron well this does allow SIGHUP | 17:58 |
| tbarron | jgriffith: right | 17:58 |
| tbarron | jgriffith: I'm agreeing that "marking" the options as safe or needing care would be good | 17:59 |
| jgriffith | oh.. got ya | 17:59 |
| tbarron | and saying that SIGHUP is not the only way to screw up with the options that aren't safe | 17:59 |
| winston-d | our fingers are much fatter than we think | 18:00 |
| tbarron | winston-d: :) | 18:00 |
| tbarron | winston-d: is it 2am or 3am where you are? | 18:01 |
| winston-d | jgriffith, smcginnis: so yeah, 'mutable' cfg is one good chnage to make if dynamic reloading is inevitable. | 18:02 |
| winston-d | tbarron: it's 2am. i shall get some sleep before my next meeting. | 18:03 |
| smcginnis | I should probably update that to include log levels so theirs at least an example using the mutable-ness. | 18:03 |
| tbarron | winston-d: i just sleep during the meetings | 18:03 |
| tbarron | smcginnis: +1 | 18:03 |
| smcginnis | *there's | 18:03 |
| hemna | could we use a special restart/reload RPC message to cut down on the window to all of the nodes? | 18:04 |
| winston-d | tbarron: i need to learn sleep talking sooner. ;) | 18:04 |
| tbarron | :) | 18:04 |
| winston-d | ok, signing off now. good day to you all. | 18:04 |
| *** winston-d is now known as winston-d_zZZ | 18:05 | |
| hemna | dunno, this seems like a lot of hand wavy to me and not enough thought/design behind it | 18:05 |
| *** jwcroppe has joined #openstack-cinder | 18:07 | |
| *** dustins_ has joined #openstack-cinder | 18:07 | |
| *** pewp has quit IRC | 18:08 | |
| *** dustins has quit IRC | 18:08 | |
| *** Apoorva_ has quit IRC | 18:09 | |
| *** Apoorva has joined #openstack-cinder | 18:10 | |
| *** pewp has joined #openstack-cinder | 18:11 | |
| *** salv-orl_ has quit IRC | 18:17 | |
| openstackgerrit | Eric Young proposed openstack/cinder master: ScaleIO: Logging cleanup https://review.openstack.org/476225 | 18:18 |
| *** stvnoyes has left #openstack-cinder | 18:19 | |
| mchlumsky | @eric_young, we disabled the Zero padding feature in SIO for performance concerns which means we can get data from previously deleted volumes. With the volume_clear feature we could quickly zero-out the first few megabytes and no longer have a filesystem. | 18:20 |
| eric_young | mchlumsky: make sense but please follow up with the normal support channels so this enhancement can get prioritized. It's possible I can support something like that in Pike+ | 18:24 |
| mchlumsky | @eric_young, will do. I'll open a bug report as well. thanks! | 18:25 |
| *** jwcroppe has quit IRC | 18:50 | |
| *** wewe0901 has quit IRC | 19:05 | |
| *** nicolasbock has quit IRC | 19:07 | |
| *** arnewiebalck_ has joined #openstack-cinder | 19:25 | |
| *** lkuchlan has joined #openstack-cinder | 19:28 | |
| *** lkuchlan has quit IRC | 19:28 | |
| *** kaisers_ has quit IRC | 19:36 | |
| *** ChubYann has joined #openstack-cinder | 19:48 | |
| *** crushil has quit IRC | 19:50 | |
| *** jwcroppe has joined #openstack-cinder | 20:00 | |
| *** crushil has joined #openstack-cinder | 20:02 | |
| *** openstackgerrit has quit IRC | 20:03 | |
| *** salv-orlando has joined #openstack-cinder | 20:04 | |
| *** kaisers_ has joined #openstack-cinder | 20:05 | |
| *** gouthamr has quit IRC | 20:09 | |
| *** liverpooler has quit IRC | 20:12 | |
| *** jwcroppe_ has joined #openstack-cinder | 20:17 | |
| *** jwcroppe has quit IRC | 20:21 | |
| *** openstackgerrit has joined #openstack-cinder | 20:22 | |
| openstackgerrit | xing-yang proposed openstack/cinder master: Functional tests for group replication https://review.openstack.org/472784 | 20:22 |
| openstackgerrit | xing-yang proposed openstack/python-cinderclient master: Tiramisu: replication group support https://review.openstack.org/352229 | 20:23 |
| *** gouthamr has joined #openstack-cinder | 20:30 | |
| *** gouthamr has quit IRC | 20:35 | |
| openstackgerrit | Chaynika Saikia proposed openstack/python-cinderclient master: Closes-Bug: 1696556 https://review.openstack.org/476264 | 20:38 |
| openstack | bug 1696556 in python-cinderclient "UnboundLocalError on message-list" [Undecided,In progress] https://launchpad.net/bugs/1696556 - Assigned to Chaynika Saikia (csaikia) | 20:38 |
| *** arnewiebalck_ has quit IRC | 20:40 | |
| *** gouthamr has joined #openstack-cinder | 20:48 | |
| *** crushil has quit IRC | 20:53 | |
| *** lucasxu has quit IRC | 20:54 | |
| *** mchlumsky has quit IRC | 21:08 | |
| *** crushil has joined #openstack-cinder | 21:09 | |
| openstackgerrit | Chaynika Saikia proposed openstack/python-cinderclient master: UnboundLocalError on message-list https://review.openstack.org/476264 | 21:09 |
| *** abishop has quit IRC | 21:13 | |
| *** chlong has quit IRC | 21:22 | |
| *** eric_young has quit IRC | 21:28 | |
| openstackgerrit | Merged openstack/python-cinderclient master: Enabled like filter support in client https://review.openstack.org/466214 | 21:37 |
| *** cknight has quit IRC | 21:49 | |
| *** catintheroof has quit IRC | 21:51 | |
| *** markvoelker has quit IRC | 22:04 | |
| *** markvoelker_ has joined #openstack-cinder | 22:06 | |
| *** markvoelker_ has quit IRC | 22:11 | |
| *** felipemonteiro_ has quit IRC | 22:12 | |
| *** salv-orl_ has joined #openstack-cinder | 22:12 | |
| *** salv-orlando has quit IRC | 22:15 | |
| *** salv-orl_ has quit IRC | 22:15 | |
| *** Julien-zte has joined #openstack-cinder | 22:15 | |
| *** jwcroppe has joined #openstack-cinder | 22:15 | |
| *** salv-orlando has joined #openstack-cinder | 22:16 | |
| *** markvoelker has joined #openstack-cinder | 22:17 | |
| *** xyang1 has quit IRC | 22:18 | |
| *** jwcroppe_ has quit IRC | 22:19 | |
| *** jwcroppe has quit IRC | 22:20 | |
| *** salv-orlando has quit IRC | 22:21 | |
| *** dustins_ has quit IRC | 22:39 | |
| *** Julien-zte has quit IRC | 22:50 | |
| *** Julien-zte has joined #openstack-cinder | 22:51 | |
| *** Apoorva_ has joined #openstack-cinder | 23:01 | |
| *** Apoorva has quit IRC | 23:05 | |
| *** sdague has quit IRC | 23:11 | |
| *** Julien-zte has quit IRC | 23:32 | |
| *** diablo_rojo has quit IRC | 23:33 | |
| *** Apoorva_ has quit IRC | 23:33 | |
| *** Apoorva has joined #openstack-cinder | 23:34 | |
| *** hoonetorg has quit IRC | 23:39 | |
| *** flwang2 has quit IRC | 23:42 | |
| *** s-shiono has joined #openstack-cinder | 23:43 | |
| *** salv-orlando has joined #openstack-cinder | 23:47 | |
| *** salv-orlando has quit IRC | 23:52 | |
| *** KeithMnemonic has joined #openstack-cinder | 23:54 | |
| *** hoonetorg has joined #openstack-cinder | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!