| openstackgerrit | Merged openstack/cinder master: Disallow modification of in use Volume Types https://review.openstack.org/440680 | 00:00 |
|---|---|---|
| *** liverpooler has quit IRC | 00:01 | |
| *** alonma has joined #openstack-cinder | 00:02 | |
| openstackgerrit | Merged openstack/cinder master: cinder-backup - coordination not initialized https://review.openstack.org/441190 | 00:02 |
| openstackgerrit | Merged openstack/cinder master: Add per-backend availability zones https://review.openstack.org/433437 | 00:02 |
| openstackgerrit | Merged openstack/cinder master: Extracted HTTP response codes to constants https://review.openstack.org/426749 | 00:02 |
| *** asselin has quit IRC | 00:04 | |
| *** alonma has quit IRC | 00:06 | |
| *** mriedem has quit IRC | 00:06 | |
| *** yangyapeng has quit IRC | 00:10 | |
| *** delatte has quit IRC | 00:13 | |
| *** _cjones_ has quit IRC | 00:21 | |
| *** _cjones_ has joined #openstack-cinder | 00:21 | |
| *** gcb has quit IRC | 00:26 | |
| *** _cjones_ has quit IRC | 00:26 | |
| *** ckonstan` has quit IRC | 00:28 | |
| *** lixiaoy1 has joined #openstack-cinder | 00:29 | |
| *** chhavi has joined #openstack-cinder | 00:40 | |
| *** zsli has joined #openstack-cinder | 00:42 | |
| *** zsli has quit IRC | 00:42 | |
| *** Shunli has joined #openstack-cinder | 00:42 | |
| *** chhavi has quit IRC | 00:44 | |
| tommylikehu | ping jgriffith, hope you still around :) | 00:46 |
| Apoorva | My Cinder CI is failing with the following error http://openstack-ci.tintri.com/tintri/refs-changes-28-352228-10/stack.sh.log.out has anyone seen this? Do I need to upgrade any dependencies? | 00:49 |
| *** mriedem has joined #openstack-cinder | 00:49 | |
| smcginnis | Apoorva: Hmm, something looks stale in your image. | 00:50 |
| smcginnis | Apoorva: I haven't seen that. I know pbr was bumped, but I haven't seen that failure anywhere else. | 00:51 |
| Apoorva | smcginnis last week I upgraded Ubuntu to 16.04 since 14.04 is no longer supported! | 00:51 |
| smcginnis | Apoorva: Yep, trusty is not supported anymore. I think you can pass a --force flag to devstack, but otherwise it will not stack. | 00:52 |
| Apoorva | smcginnis I has successful runs on 16.04 but things started failng after that | 00:52 |
| Apoorva | had** | 00:52 |
| smcginnis | Apoorva: Do you pre-stage instances? Maybe something is left over that is no longer compatible. | 00:53 |
| *** salv-orlando has joined #openstack-cinder | 00:53 | |
| Apoorva | smcginnis is there any way I can force clean and fresh install? | 00:54 |
| *** ducttape_ has joined #openstack-cinder | 00:54 | |
| smcginnis | Apoorva: Well, I would just use a fresh 16.04 image, but for one that has already run devstack you can try running the clean.sh script in the root of the devstack repo. | 00:55 |
| smcginnis | Apoorva: After that you could also try "sudo apt autoremove" to try to clean out other packages. | 00:56 |
| Apoorva | smcginnis thanks. I will try that | 00:56 |
| *** ducttape_ has quit IRC | 00:56 | |
| *** ducttape_ has joined #openstack-cinder | 00:57 | |
| smcginnis | Apoorva: Good luck! | 00:57 |
| *** salv-orlando has quit IRC | 00:58 | |
| *** sdague has quit IRC | 01:01 | |
| *** zsli_ has joined #openstack-cinder | 01:01 | |
| *** namnh has joined #openstack-cinder | 01:03 | |
| *** jiaopengju has joined #openstack-cinder | 01:03 | |
| *** diablo_rojo has quit IRC | 01:04 | |
| *** Shunli has quit IRC | 01:04 | |
| *** zhurong has joined #openstack-cinder | 01:06 | |
| *** ducttape_ has quit IRC | 01:08 | |
| *** kongwei has joined #openstack-cinder | 01:08 | |
| *** ntpttr_ has quit IRC | 01:11 | |
| *** yangyapeng has joined #openstack-cinder | 01:13 | |
| *** yangyapeng has quit IRC | 01:14 | |
| *** yangyapeng has joined #openstack-cinder | 01:14 | |
| tommylikehu | ping smcginnis | 01:14 |
| tommylikehu | smcginnis: I noticed you and jgriffith's chat on filtering resource on a more generic way, what about the other filter that doesn't related to filtering by database column. | 01:17 |
| tommylikehu | this one: https://review.openstack.org/#/c/362747/. | 01:17 |
| jgriffith | tommylikehu hey | 01:18 |
| *** kongwei has quit IRC | 01:18 | |
| tommylikehu | hey jgriffith | 01:18 |
| jgriffith | tommylikehu so the things the fk tables is another issue that I haven't completely decided on yet | 01:18 |
| jgriffith | tommylikehu you'll notice the get object and the volume object table itself has the ability to populate/use these filters | 01:19 |
| jgriffith | tommylikehu but like I said in the review, hopefully we'll NOT be needing things like attachment info much longer | 01:19 |
| jgriffith | tommylikehu the instance-uuid is easy since we already have that info | 01:19 |
| jgriffith | tommylikehu things like host shouldn't leak out to the users anyway IMHO | 01:20 |
| tommylikehu | jgriffith: I agree with you on the attachment spec:) | 01:20 |
| jgriffith | We've sort of blurred the lines between user, admin and cloud-admin | 01:20 |
| jgriffith | tommylikehu so you can filter on types already FYI | 01:21 |
| tommylikehu | jgriffith: but I am collecting more reviews for this spec: https://review.openstack.org/#/c/362747/ this filter in this spec will be used at the cinder-scheduler | 01:21 |
| tommylikehu | what do you think of this :) | 01:21 |
| jgriffith | ahhh | 01:21 |
| jgriffith | lemme look | 01:22 |
| *** mayela has quit IRC | 01:22 | |
| jgriffith | tommylikehu so I'll have to look at it more closely, but in general I think adding the ability isn't a bad idea at all. | 01:23 |
| jgriffith | tommylikehu I would however prefer we just add a --filter or --search-options to the cinderclient command and avoid adding the one-offs in the api constantly | 01:23 |
| jgriffith | tommylikehu does that make sense? | 01:23 |
| tommylikehu | thanks, so what are you proposing is more related to expose database column and refactor client's argument like list --filter abc=xyz? | 01:24 |
| jgriffith | tommylikehu correct, in fact I'm hopefully going to push that up tonight | 01:24 |
| jgriffith | for volumes | 01:24 |
| tommylikehu | jgriffith: that sounds absolutely great ! | 01:24 |
| jgriffith | then once we all settle on that we can look at the other objects | 01:25 |
| jgriffith | :) | 01:25 |
| jgriffith | cool, glad you like the idea | 01:25 |
| jgriffith | I just figured it would be nice to keep the command switches minimized and give more flexibility in filters going forward | 01:25 |
| tommylikehu | jgriffith:yes, one more question, about the filters | 01:25 |
| jgriffith | and if we can make it consistent across all objects that will be a huge help | 01:25 |
| jgriffith | tommylikehu sure? | 01:26 |
| tommylikehu | what if the filter value is a json not a sure? | 01:26 |
| jgriffith | you mean json vs k=v format? | 01:26 |
| *** Apoorva has quit IRC | 01:26 | |
| tommylikehu | yes, maybe filter a.b = 123 | 01:26 |
| jgriffith | from the API servers view it's json anyway; but from the client you want to specify json? | 01:26 |
| jgriffith | oh... maybe you haven't seen this: https://gist.github.com/j-griffith/f455e51d3e597af96bfe7022974b5bd9 | 01:27 |
| jgriffith | or do you mean something different? | 01:28 |
| *** lucasxu has joined #openstack-cinder | 01:28 | |
| jgriffith | of course we would make that so you can pass in multiple filters | 01:28 |
| tommylikehu | jgriffith: I mean --filter image_metadata.key1 = 123 something like that | 01:29 |
| jgriffith | tommylikehu Oh! Yeah, that's sort of what I had in mind for the next step to do relational filters | 01:32 |
| jgriffith | but honestly I'm not sure how I feel about those yet | 01:32 |
| *** ducttape_ has joined #openstack-cinder | 01:32 | |
| jgriffith | I'd like to get a change with the base object columns first, then work on the relationships | 01:33 |
| tommylikehu | jgriffith: thanks, maybe my suggestion only looks right at first sight, | 01:34 |
| jgriffith | tommylikehu but your suggestion of "<table>.<key> = value" is IMO probably the best way to add those | 01:34 |
| jgriffith | tommylikehu it gets a little tricky/ugly in setting up the config settings though... so we need to think about that a bit | 01:35 |
| tommylikehu | config of what are validate columns? | 01:35 |
| jgriffith | tommylikehu https://github.com/openstack/cinder/blob/master/cinder/api/common.py#L42 | 01:36 |
| jgriffith | tommylikehu so I stumbled across that today while building something else :) | 01:36 |
| jgriffith | tommylikehu so that would allow the cloud-admin to select exactly what he/she wants to expose to users | 01:37 |
| tommylikehu | jgriffith: it would add a lot of burden as you are going to support most of the resources and their columns are different | 01:38 |
| jgriffith | tommylikehu yeah, I've been thinking about that.... | 01:38 |
| jgriffith | tommylikehu I'm considering pointing to a json file | 01:38 |
| jgriffith | and generating a default json file to be used. | 01:38 |
| tommylikehu | how about add a new configure option which can specify the resource name | 01:38 |
| jgriffith | Keep in mind though.. this is "regular user" visible, so in theory the list shouldn't be that large | 01:39 |
| jgriffith | tommylikehu not sure.. can you elaborate? | 01:39 |
| tommylikehu | --query_resource_filters | 01:39 |
| jgriffith | tommylikehu you mean like just turn xyz resources on or off? | 01:39 |
| tommylikehu | no | 01:39 |
| tommylikehu | volumes.name/volumes.status something like that | 01:40 |
| jgriffith | hmm.... still not sure I'm following; but here's what I was thinking: | 01:40 |
| jgriffith | {volumes: status, volumes: name.....} | 01:41 |
| jgriffith | {snapshots: status, snapshots: name....} | 01:41 |
| tommylikehu | --query_resource_filter [volumes.name, snapshot.name] | 01:41 |
| jgriffith | but maybe something like {volume_filters: [name, type...], snapshot_filters: [status, source_volume_id...]} | 01:42 |
| tommylikehu | jgriffith: yeah | 01:43 |
| jgriffith | tommylikehu ok, for sure.. and I'd like to provide an API call to query that... is that the other part you were thinking about? | 01:43 |
| tommylikehu | provide a API call to collect the config options? | 01:44 |
| jgriffith | tommylikehu well just the filters in particular: `cinder filters-list`. dumps them all and `cinder filters-list --resource <resource-type>` for a specific resource | 01:45 |
| jgriffith | tommylikehu the main thing for me is that right now the API is kinda silly, because we have specific arguments and micro-versions to just explicitly pass in the filter arguments that just go straight to the DB anyway | 01:46 |
| jgriffith | tommylikehu that's mostly fixed in the server side, just needs worked up in the client side to start | 01:46 |
| tommylikehu | jgriffith: great~ | 01:47 |
| jgriffith | I'll ping you as soon as I get the first round of things posted | 01:47 |
| *** ducttape_ has quit IRC | 01:47 | |
| jgriffith | I'd greatly appreciate your feedback | 01:47 |
| tommylikehu | jgriffith: no problem | 01:48 |
| *** wanghao has joined #openstack-cinder | 01:50 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Add filter, sorter for group snapshot https://review.openstack.org/441651 | 01:52 |
| *** salv-orlando has joined #openstack-cinder | 01:55 | |
| *** salv-orlando has quit IRC | 01:59 | |
| *** edmondsw has joined #openstack-cinder | 02:00 | |
| *** edmondsw has quit IRC | 02:00 | |
| *** edmondsw has joined #openstack-cinder | 02:00 | |
| *** Apoorva has joined #openstack-cinder | 02:00 | |
| *** zsli__ has joined #openstack-cinder | 02:01 | |
| *** zsli_ has quit IRC | 02:03 | |
| *** mtanino_ has quit IRC | 02:09 | |
| *** erlon has quit IRC | 02:15 | |
| *** markvoelker has quit IRC | 02:16 | |
| *** ducttape_ has joined #openstack-cinder | 02:16 | |
| *** ducttape_ has quit IRC | 02:18 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Fix generic group's unit testcase https://review.openstack.org/442236 | 02:23 |
| *** akapil has joined #openstack-cinder | 02:40 | |
| *** akapil has quit IRC | 02:44 | |
| *** alonma has joined #openstack-cinder | 02:45 | |
| *** alonma has quit IRC | 02:49 | |
| *** dhinesh has quit IRC | 02:50 | |
| *** Apoorva has quit IRC | 02:50 | |
| *** edmondsw has quit IRC | 02:55 | |
| *** edmondsw has joined #openstack-cinder | 02:56 | |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/cinder master: Switch to decorators.idempotent_id https://review.openstack.org/442239 | 02:58 |
| *** wanghao has quit IRC | 02:58 | |
| *** wanghao has joined #openstack-cinder | 02:58 | |
| *** ducttape_ has joined #openstack-cinder | 03:00 | |
| *** edmondsw has quit IRC | 03:01 | |
| openstackgerrit | wangxiyuan proposed openstack/cinder master: [WIP]Don't change volume's status when create backups from snapshots https://review.openstack.org/442241 | 03:01 |
| *** bkopilov has quit IRC | 03:11 | |
| *** tlian has quit IRC | 03:14 | |
| *** ducttape_ has quit IRC | 03:14 | |
| *** lucasxu has quit IRC | 03:15 | |
| *** sudipto has joined #openstack-cinder | 03:15 | |
| *** sudipto_ has joined #openstack-cinder | 03:15 | |
| openstackgerrit | TommyLike proposed openstack/cinder master: Fix generic group's unit testcase https://review.openstack.org/442236 | 03:16 |
| *** lucasxu has joined #openstack-cinder | 03:24 | |
| *** lucasxu has quit IRC | 03:24 | |
| *** wanghao_ has joined #openstack-cinder | 03:25 | |
| *** bswartz has quit IRC | 03:25 | |
| *** wanghao has quit IRC | 03:29 | |
| *** Apoorva has joined #openstack-cinder | 03:32 | |
| *** tlian has joined #openstack-cinder | 03:33 | |
| openstackgerrit | TommyLike proposed openstack/cinder-specs master: Cinder volume revert to snapshot https://review.openstack.org/316540 | 03:34 |
| *** links has joined #openstack-cinder | 03:35 | |
| openstackgerrit | xing-yang proposed openstack/cinder master: Tiramisu: replication group support https://review.openstack.org/352228 | 03:40 |
| tommylikehu | hey anyone would like to review this patch that fix the testcases https://review.openstack.org/442236 ? It's simple and esay to review :) | 03:41 |
| openstackgerrit | xing-yang proposed openstack/cinder master: Tiramisu: replication group support https://review.openstack.org/352228 | 03:52 |
| *** gouthamr has quit IRC | 03:55 | |
| *** salv-orlando has joined #openstack-cinder | 03:55 | |
| *** mriedem has quit IRC | 03:56 | |
| *** alonma has joined #openstack-cinder | 03:57 | |
| *** salv-orlando has quit IRC | 04:00 | |
| *** Apoorva_ has joined #openstack-cinder | 04:01 | |
| *** Apoorva has quit IRC | 04:01 | |
| *** alonma has quit IRC | 04:01 | |
| *** chenhb has quit IRC | 04:10 | |
| *** chenhb has joined #openstack-cinder | 04:10 | |
| openstackgerrit | xing-yang proposed openstack/python-cinderclient master: Tiramisu: replication group support https://review.openstack.org/352229 | 04:12 |
| *** ducttape_ has joined #openstack-cinder | 04:15 | |
| *** markvoelker has joined #openstack-cinder | 04:16 | |
| *** chhavi has joined #openstack-cinder | 04:17 | |
| *** ducttape_ has quit IRC | 04:20 | |
| *** markvoelker has quit IRC | 04:21 | |
| *** psachin has joined #openstack-cinder | 04:25 | |
| *** dhinesh has joined #openstack-cinder | 04:30 | |
| *** sudipto has quit IRC | 04:32 | |
| *** sudipto_ has quit IRC | 04:32 | |
| *** sudipto has joined #openstack-cinder | 04:33 | |
| *** sudipto_ has joined #openstack-cinder | 04:33 | |
| *** pdeore has joined #openstack-cinder | 04:33 | |
| *** bkopilov has joined #openstack-cinder | 04:34 | |
| *** Apoorva_ has quit IRC | 04:35 | |
| *** bswartz has joined #openstack-cinder | 04:36 | |
| *** nicolasbock has quit IRC | 04:38 | |
| *** akapil has joined #openstack-cinder | 04:38 | |
| *** akapil has quit IRC | 04:42 | |
| openstackgerrit | Soffie Huang proposed openstack/cinder master: Extend Cinder driver to utilize multiple FSS storage pools https://review.openstack.org/422462 | 04:43 |
| *** sudipto_ has quit IRC | 04:43 | |
| *** sudipto has quit IRC | 04:43 | |
| *** mdnadeem has joined #openstack-cinder | 04:55 | |
| *** Apoorva has joined #openstack-cinder | 04:56 | |
| *** Poornima has joined #openstack-cinder | 05:10 | |
| *** ducttape_ has joined #openstack-cinder | 05:16 | |
| *** ducttape_ has quit IRC | 05:21 | |
| openstackgerrit | Merged openstack/cinder master: Fix generic group's unit testcase https://review.openstack.org/442236 | 05:21 |
| *** wanghao_ has quit IRC | 05:36 | |
| *** wanghao has joined #openstack-cinder | 05:36 | |
| *** rcernin has joined #openstack-cinder | 05:39 | |
| *** salv-orlando has joined #openstack-cinder | 05:39 | |
| *** sudipto has joined #openstack-cinder | 05:39 | |
| *** sudipto_ has joined #openstack-cinder | 05:39 | |
| *** salv-orlando has quit IRC | 05:43 | |
| *** Jack_I has joined #openstack-cinder | 05:44 | |
| *** markvoelker has joined #openstack-cinder | 05:46 | |
| *** tlian has quit IRC | 05:47 | |
| *** markvoelker has quit IRC | 05:50 | |
| *** wanghao has quit IRC | 05:54 | |
| *** wanghao has joined #openstack-cinder | 05:55 | |
| *** alonmarx has joined #openstack-cinder | 05:55 | |
| *** Jack_Iv has joined #openstack-cinder | 05:55 | |
| *** Jack_Iv has quit IRC | 05:57 | |
| *** zsli__ has quit IRC | 05:58 | |
| *** zsli__ has joined #openstack-cinder | 05:59 | |
| *** pgadiya has joined #openstack-cinder | 05:59 | |
| *** Jack_Iv has joined #openstack-cinder | 06:00 | |
| *** yangyape_ has joined #openstack-cinder | 06:01 | |
| *** wanghao has quit IRC | 06:02 | |
| *** wanghao has joined #openstack-cinder | 06:02 | |
| *** yangyapeng has quit IRC | 06:03 | |
| *** wanghao_ has joined #openstack-cinder | 06:07 | |
| *** wanghao has quit IRC | 06:11 | |
| *** yangyapeng has joined #openstack-cinder | 06:12 | |
| *** Apoorva has quit IRC | 06:12 | |
| *** rcernin has quit IRC | 06:13 | |
| *** Jack_Iv has quit IRC | 06:15 | |
| *** yangyape_ has quit IRC | 06:15 | |
| *** yangyape_ has joined #openstack-cinder | 06:16 | |
| *** salv-orlando has joined #openstack-cinder | 06:16 | |
| *** ducttape_ has joined #openstack-cinder | 06:17 | |
| *** adisky_ has joined #openstack-cinder | 06:19 | |
| *** yangyapeng has quit IRC | 06:20 | |
| *** yangyape_ has quit IRC | 06:21 | |
| *** Jack_Iv has joined #openstack-cinder | 06:22 | |
| *** ducttape_ has quit IRC | 06:22 | |
| *** yangyapeng has joined #openstack-cinder | 06:23 | |
| openstackgerrit | jiamin proposed openstack/cinder master: DS8K: specify pool and lss in extra-specs https://review.openstack.org/435857 | 06:25 |
| *** gcb has joined #openstack-cinder | 06:28 | |
| *** ChubYann has quit IRC | 06:33 | |
| openstackgerrit | Rick Chen proposed openstack/cinder master: HTTPConnect SSL certification verification failed https://review.openstack.org/442288 | 06:33 |
| *** e0ne has joined #openstack-cinder | 06:36 | |
| *** markvoelker has joined #openstack-cinder | 06:36 | |
| *** Jack_Iv has quit IRC | 06:37 | |
| *** savihou has joined #openstack-cinder | 06:39 | |
| *** nkrinner_afk is now known as nkrinner | 06:40 | |
| *** e0ne has quit IRC | 06:40 | |
| *** namnh has quit IRC | 06:48 | |
| *** markvoelker has quit IRC | 06:49 | |
| *** lamt has joined #openstack-cinder | 06:51 | |
| *** zsli__ has quit IRC | 06:56 | |
| *** zsli__ has joined #openstack-cinder | 06:57 | |
| *** e0ne has joined #openstack-cinder | 06:57 | |
| *** lpetrut has joined #openstack-cinder | 06:58 | |
| *** lamt has quit IRC | 07:06 | |
| *** zsli__ has quit IRC | 07:08 | |
| *** e0ne has quit IRC | 07:15 | |
| *** lkuchlan has joined #openstack-cinder | 07:18 | |
| *** ducttape_ has joined #openstack-cinder | 07:18 | |
| *** salv-orl_ has joined #openstack-cinder | 07:23 | |
| *** ducttape_ has quit IRC | 07:23 | |
| *** salv-orlando has quit IRC | 07:26 | |
| *** namnh has joined #openstack-cinder | 07:35 | |
| *** salv-orl_ has quit IRC | 07:35 | |
| *** lpetrut has quit IRC | 07:43 | |
| *** gcb has quit IRC | 07:45 | |
| *** tesseract has joined #openstack-cinder | 07:47 | |
| *** alonmarx has quit IRC | 07:49 | |
| *** pcaruana has joined #openstack-cinder | 07:50 | |
| *** dhinesh has quit IRC | 07:50 | |
| *** isaacb has joined #openstack-cinder | 07:52 | |
| *** salv-orlando has joined #openstack-cinder | 07:53 | |
| *** rcernin has joined #openstack-cinder | 07:55 | |
| openstackgerrit | Soffie Huang proposed openstack/cinder master: Extend Cinder driver to utilize multiple FSS storage pools https://review.openstack.org/422462 | 07:57 |
| *** dalgaaf has quit IRC | 07:57 | |
| *** dalgaaf has joined #openstack-cinder | 07:57 | |
| *** dosaboy has quit IRC | 07:58 | |
| *** gcb has joined #openstack-cinder | 07:58 | |
| *** dosaboy has joined #openstack-cinder | 07:59 | |
| *** akapil has joined #openstack-cinder | 08:00 | |
| openstackgerrit | wangxiyuan proposed openstack/cinder master: Don't change volume's status when create backups from snapshots https://review.openstack.org/442241 | 08:02 |
| *** akapil has quit IRC | 08:05 | |
| *** akapil has joined #openstack-cinder | 08:05 | |
| *** alonma has joined #openstack-cinder | 08:10 | |
| *** lpetrut has joined #openstack-cinder | 08:12 | |
| *** yangyape_ has joined #openstack-cinder | 08:19 | |
| *** yangyapeng has quit IRC | 08:22 | |
| *** s-shiono has quit IRC | 08:40 | |
| *** sam_nowitzki has joined #openstack-cinder | 08:49 | |
| openstackgerrit | jiaohaolin proposed openstack/cinder master: Inspur Cinder FC/iSCSI driver https://review.openstack.org/439488 | 08:51 |
| *** lixiaoy1 has quit IRC | 08:52 | |
| openstackgerrit | chenhb proposed openstack/cinder master: Fix the missing args bug while lvm creates volume failed https://review.openstack.org/442322 | 08:59 |
| *** zzzeek has quit IRC | 09:00 | |
| *** pgadiya has quit IRC | 09:01 | |
| *** zzzeek has joined #openstack-cinder | 09:01 | |
| *** chhavi has quit IRC | 09:05 | |
| *** pgadiya has joined #openstack-cinder | 09:06 | |
| *** akapil has quit IRC | 09:14 | |
| *** lkuchlan has quit IRC | 09:14 | |
| *** lpetrut has quit IRC | 09:15 | |
| *** akapil has joined #openstack-cinder | 09:15 | |
| *** wanghao_ has quit IRC | 09:17 | |
| *** lpetrut has joined #openstack-cinder | 09:17 | |
| *** ducttape_ has joined #openstack-cinder | 09:19 | |
| *** mvk has quit IRC | 09:21 | |
| *** faizy has joined #openstack-cinder | 09:22 | |
| *** wanghao has joined #openstack-cinder | 09:23 | |
| *** belmoreira has joined #openstack-cinder | 09:24 | |
| *** lkuchlan has joined #openstack-cinder | 09:24 | |
| *** ducttape_ has quit IRC | 09:25 | |
| *** karthikp_ has quit IRC | 09:25 | |
| *** bkopilov has quit IRC | 09:26 | |
| *** wanghao has quit IRC | 09:27 | |
| *** bkopilov has joined #openstack-cinder | 09:29 | |
| *** karthikp_ has joined #openstack-cinder | 09:31 | |
| *** chhavi has joined #openstack-cinder | 09:31 | |
| *** Poornima has quit IRC | 09:35 | |
| *** redondo-mk has joined #openstack-cinder | 09:41 | |
| *** Jack_Iv has joined #openstack-cinder | 09:45 | |
| openstackgerrit | Merged openstack/cinder master: Fix action on services raises TypeError https://review.openstack.org/439403 | 09:45 |
| *** lkuchlan has quit IRC | 09:46 | |
| *** aarefiev_afk is now known as aarefiev | 09:47 | |
| *** Jack_Iv has quit IRC | 09:47 | |
| *** mvk has joined #openstack-cinder | 09:52 | |
| *** lkuchlan has joined #openstack-cinder | 09:55 | |
| *** bkopilov has quit IRC | 09:56 | |
| *** e0ne has joined #openstack-cinder | 10:03 | |
| *** links has quit IRC | 10:05 | |
| *** Poornima has joined #openstack-cinder | 10:06 | |
| *** bkopilov has joined #openstack-cinder | 10:09 | |
| *** markvoelker has joined #openstack-cinder | 10:10 | |
| *** ociuhandu has joined #openstack-cinder | 10:10 | |
| *** namnh_ has joined #openstack-cinder | 10:10 | |
| *** e0ne has quit IRC | 10:12 | |
| *** namnh has quit IRC | 10:13 | |
| *** yangyape_ has quit IRC | 10:14 | |
| *** namnh_ has quit IRC | 10:16 | |
| openstackgerrit | jiamin proposed openstack/cinder master: DS8K: specify pool and lss in extra-specs https://review.openstack.org/435857 | 10:17 |
| *** links has joined #openstack-cinder | 10:19 | |
| *** ducttape_ has joined #openstack-cinder | 10:21 | |
| *** alonma has quit IRC | 10:21 | |
| *** alonma has joined #openstack-cinder | 10:22 | |
| *** salv-orl_ has joined #openstack-cinder | 10:23 | |
| *** e0ne has joined #openstack-cinder | 10:24 | |
| *** ducttape_ has quit IRC | 10:26 | |
| *** gcb has quit IRC | 10:26 | |
| *** alonma has quit IRC | 10:26 | |
| *** salv-orlando has quit IRC | 10:27 | |
| *** salv-orl_ has quit IRC | 10:27 | |
| *** mvk has quit IRC | 10:28 | |
| *** Poornima has quit IRC | 10:30 | |
| *** Poornima has joined #openstack-cinder | 10:30 | |
| *** Poornima has quit IRC | 10:31 | |
| *** Poornima has joined #openstack-cinder | 10:32 | |
| *** alonma has joined #openstack-cinder | 10:33 | |
| *** openstackgerrit has quit IRC | 10:33 | |
| *** alonma has quit IRC | 10:37 | |
| *** links has quit IRC | 10:39 | |
| *** sam_nowitzki has quit IRC | 10:39 | |
| *** dalgaaf has quit IRC | 10:40 | |
| *** mvk has joined #openstack-cinder | 10:41 | |
| *** Guest6667 is now known as Guest6666 | 10:43 | |
| *** dalgaaf has joined #openstack-cinder | 10:44 | |
| *** alonma has joined #openstack-cinder | 10:44 | |
| *** akapil has quit IRC | 10:45 | |
| *** alonma has quit IRC | 10:48 | |
| *** isaacb has quit IRC | 10:50 | |
| *** pdeore has quit IRC | 10:55 | |
| *** alonma has joined #openstack-cinder | 10:55 | |
| *** pdeore has joined #openstack-cinder | 10:55 | |
| *** links has joined #openstack-cinder | 10:55 | |
| *** akapil has joined #openstack-cinder | 10:55 | |
| *** alonma has quit IRC | 10:58 | |
| *** alonma has joined #openstack-cinder | 10:58 | |
| *** akapil has quit IRC | 11:00 | |
| *** akapil has joined #openstack-cinder | 11:05 | |
| *** Poornima has quit IRC | 11:05 | |
| *** akapil has quit IRC | 11:06 | |
| *** akapil has joined #openstack-cinder | 11:06 | |
| *** chhavi has quit IRC | 11:13 | |
| *** jiaopengju has quit IRC | 11:15 | |
| *** edmondsw has joined #openstack-cinder | 11:19 | |
| *** chhavi has joined #openstack-cinder | 11:19 | |
| *** alonma has quit IRC | 11:19 | |
| *** alonma has joined #openstack-cinder | 11:20 | |
| *** nicolasbock has joined #openstack-cinder | 11:21 | |
| *** edmondsw has quit IRC | 11:23 | |
| *** alonma has quit IRC | 11:24 | |
| *** sdague has joined #openstack-cinder | 11:27 | |
| *** alonma has joined #openstack-cinder | 11:31 | |
| *** alonma has quit IRC | 11:31 | |
| *** alonma has joined #openstack-cinder | 11:32 | |
| *** Poornima has joined #openstack-cinder | 11:33 | |
| *** isaacb has joined #openstack-cinder | 11:34 | |
| *** sudipto has quit IRC | 11:36 | |
| *** sudipto_ has quit IRC | 11:36 | |
| *** yangyapeng has joined #openstack-cinder | 11:39 | |
| *** gkadam has joined #openstack-cinder | 11:40 | |
| *** yangyapeng has quit IRC | 11:40 | |
| *** yangyapeng has joined #openstack-cinder | 11:40 | |
| gkadam | hi , is cinder v1 api deprecated? what is the status? I want to run the tempest test for a cloud env having ceph backend. so i am confused as to , if api_v1 = true and api_v2 = true needed in [volume-feature-enabled] section | 11:41 |
| *** wanghao has joined #openstack-cinder | 11:42 | |
| gkadam | in tempest.conf | 11:42 |
| e0ne | gkadam: hi. both v1 and v2 APIs are deprecated now | 11:44 |
| smcginnis | gkadam: Hi. Yes, the v1 API has been deprecated for some time now and is now defaulted to be off. | 11:44 |
| links | gkadam, https://developer.openstack.org/api-ref/block-storage/ | 11:50 |
| *** markvoelker has quit IRC | 11:50 | |
| gkadam | links, smcginnis, e0ne thanks for the prompt reply. so if i understand correct, these should be set to false or not mentioned at all in tempest.conf right? | 11:52 |
| gkadam | this is the error i am getting for both the apis : | 11:53 |
| gkadam | Captured traceback: | 11:53 |
| gkadam | ~~~~~~~~~~~~~~~~~~~ | 11:53 |
| gkadam | Traceback (most recent call last): | 11:53 |
| gkadam | File "/tempest/tempest/api/volume/admin/test_volume_types.py", line 66, in test_volume_crud_with_volume_type_and_extra_specs | 11:53 |
| gkadam | volume['id'], 'available') | 11:53 |
| gkadam | File "/tempest/tempest/common/waiters.py", line 180, in wait_for_volume_status | 11:53 |
| gkadam | raise exceptions.VolumeBuildErrorException(volume_id=volume_id) | 11:53 |
| gkadam | tempest.exceptions.VolumeBuildErrorException: Volume f58e1c60-6096-41fc-9449-30cac0d31eb1 failed to build and is in ERROR status | 11:53 |
| smcginnis | gkadam: v1 should definitely be set to false. | 11:54 |
| smcginnis | gkadam: For now, I believe v2 defaults to true. | 11:54 |
| gkadam | smcginnis, thanks. i will try that! | 11:54 |
| smcginnis | gkadam: You'll have to look in your logs as far as figuring out why you are not able to create that volume. | 11:55 |
| gkadam | smcginnis, ack | 11:56 |
| *** openstackgerrit has joined #openstack-cinder | 11:58 | |
| openstackgerrit | wangxiyuan proposed openstack/cinder master: Support sort snapshots with name https://review.openstack.org/431488 | 11:58 |
| openstackgerrit | Andrey Volkov proposed openstack/cinder master: api-ref cleanup: remove quotas user related calls https://review.openstack.org/406001 | 11:59 |
| openstackgerrit | Andrey Volkov proposed openstack/cinder master: api-ref cleanup: required flag for tenants https://review.openstack.org/406056 | 11:59 |
| *** wanghao has quit IRC | 12:07 | |
| *** salv-orlando has joined #openstack-cinder | 12:09 | |
| *** faizy has quit IRC | 12:12 | |
| *** bkopilov has quit IRC | 12:13 | |
| *** isaacb has quit IRC | 12:20 | |
| *** dgonzalez has quit IRC | 12:22 | |
| *** markvoelker has joined #openstack-cinder | 12:23 | |
| *** catintheroof has joined #openstack-cinder | 12:24 | |
| *** wanghao has joined #openstack-cinder | 12:27 | |
| *** salv-orlando has quit IRC | 12:29 | |
| *** pdeore has quit IRC | 12:33 | |
| openstackgerrit | Vipin Balachandran proposed openstack/cinder master: VMware: Refactor vmdk unit tests https://review.openstack.org/442423 | 12:35 |
| *** wanghao has quit IRC | 12:37 | |
| *** wanghao has joined #openstack-cinder | 12:40 | |
| *** erlon has joined #openstack-cinder | 12:43 | |
| *** zhurong has quit IRC | 12:45 | |
| openstackgerrit | Merged openstack/cinder master: NetApp cDOT: Add check if copyoffload tool is available https://review.openstack.org/427613 | 12:50 |
| *** isaacb has joined #openstack-cinder | 12:50 | |
| *** nikeshm has joined #openstack-cinder | 12:55 | |
| *** dave-mccowan has joined #openstack-cinder | 12:55 | |
| *** Poornima has quit IRC | 12:56 | |
| *** edmondsw has joined #openstack-cinder | 13:00 | |
| *** chlong has joined #openstack-cinder | 13:00 | |
| *** porrua has joined #openstack-cinder | 13:01 | |
| *** faizy has joined #openstack-cinder | 13:01 | |
| *** dustins has joined #openstack-cinder | 13:02 | |
| *** porrua_ has joined #openstack-cinder | 13:02 | |
| *** esker has joined #openstack-cinder | 13:03 | |
| *** wanghao has quit IRC | 13:05 | |
| *** porrua has quit IRC | 13:06 | |
| *** edmondsw has quit IRC | 13:06 | |
| *** lkuchlan has quit IRC | 13:06 | |
| *** diegows has quit IRC | 13:07 | |
| *** esker has quit IRC | 13:08 | |
| *** edmondsw has joined #openstack-cinder | 13:08 | |
| *** sudipto has joined #openstack-cinder | 13:09 | |
| *** sudipto_ has joined #openstack-cinder | 13:09 | |
| *** eharney has quit IRC | 13:12 | |
| *** edmondsw has quit IRC | 13:12 | |
| *** tushar_ has joined #openstack-cinder | 13:13 | |
| tushar_ | Hi | 13:13 |
| *** isaacb has quit IRC | 13:14 | |
| *** edmondsw has joined #openstack-cinder | 13:14 | |
| *** edmondsw has quit IRC | 13:14 | |
| tushar_ | We are facing some issue while running cinder testcases in CI with iSCSI connectivity....with FC its running properly | 13:15 |
| *** edmondsw has joined #openstack-cinder | 13:15 | |
| tushar_ | Is there any changed required in the configuration ...We are running on ubuntu xenial | 13:15 |
| tushar_ | ? | 13:15 |
| smcginnis | tushar_: What kind of errors are you getting? | 13:17 |
| tushar_ | smcginnis : Cinder tescases are failing. | 13:18 |
| smcginnis | :? | 13:18 |
| *** isaacb has joined #openstack-cinder | 13:19 | |
| tushar_ | smcginnis : the error message which I saw is - FailedISCSITargetPortalLogin: Could not login to any iSCSI portal...Earlier CI was running on Trusty | 13:20 |
| smcginnis | tushar_: Well, you'll have to look through your logs to see what it was trying to do and figure out why it can't login. | 13:20 |
| *** faizy has quit IRC | 13:20 | |
| smcginnis | tushar_: Sounds like maybe the wrong IP is being returned or something. Should be easy enough to try manually to verify. | 13:21 |
| *** liverpooler has joined #openstack-cinder | 13:23 | |
| tushar_ | smcginnis : Ok. I was looking in configuration part because it starts failing after moving from Tusty to Xenial | 13:23 |
| *** ducttape_ has joined #openstack-cinder | 13:23 | |
| *** bkopilov has joined #openstack-cinder | 13:24 | |
| tushar_ | smcginnis : and the strange thing is Cinder ci with FC connectivity is running with no issues | 13:25 |
| *** Poornima has joined #openstack-cinder | 13:25 | |
| *** alonma has quit IRC | 13:25 | |
| smcginnis | tushar_: Well, since the error is that it can't log in to the iSCSI portal, that kind of makes sense. | 13:25 |
| *** alonma has joined #openstack-cinder | 13:26 | |
| *** salv-orlando has joined #openstack-cinder | 13:28 | |
| *** ducttape_ has quit IRC | 13:28 | |
| *** Jack_Iv has joined #openstack-cinder | 13:28 | |
| tushar_ | smcginnis : ok. | 13:29 |
| *** bkopilov has quit IRC | 13:29 | |
| *** wN has quit IRC | 13:30 | |
| *** alonma has quit IRC | 13:30 | |
| *** alonma has joined #openstack-cinder | 13:32 | |
| *** zhurong has joined #openstack-cinder | 13:35 | |
| *** salv-orl_ has joined #openstack-cinder | 13:36 | |
| *** pgadiya has quit IRC | 13:36 | |
| *** tommylikehu_ has joined #openstack-cinder | 13:37 | |
| *** salv-orlando has quit IRC | 13:39 | |
| *** bkopilov has joined #openstack-cinder | 13:41 | |
| *** gouthamr has joined #openstack-cinder | 13:43 | |
| *** salv-orl_ has quit IRC | 13:46 | |
| *** ducttape_ has joined #openstack-cinder | 13:49 | |
| *** lpetrut has quit IRC | 13:49 | |
| *** links has quit IRC | 13:51 | |
| *** sudipto_ has quit IRC | 13:54 | |
| *** sudipto has quit IRC | 13:54 | |
| *** crushil has joined #openstack-cinder | 13:55 | |
| openstackgerrit | Merged openstack/cinder master: Switch to decorators.idempotent_id https://review.openstack.org/442239 | 13:56 |
| *** gouthamr has quit IRC | 13:59 | |
| *** gouthamr has joined #openstack-cinder | 14:02 | |
| openstackgerrit | kavitha h r proposed openstack/os-brick master: The python 3.5 is added https://review.openstack.org/442507 | 14:03 |
| *** mdnadeem has quit IRC | 14:06 | |
| *** lpetrut has joined #openstack-cinder | 14:09 | |
| *** ducttape_ has quit IRC | 14:12 | |
| *** Poornima has quit IRC | 14:12 | |
| *** lamt has joined #openstack-cinder | 14:14 | |
| *** xyang1 has joined #openstack-cinder | 14:15 | |
| *** zhurong has quit IRC | 14:16 | |
| *** pdeore has joined #openstack-cinder | 14:17 | |
| *** nikeshm has quit IRC | 14:17 | |
| *** pdeore has quit IRC | 14:18 | |
| *** pdeore has joined #openstack-cinder | 14:18 | |
| *** ociuhandu has quit IRC | 14:18 | |
| *** laurent\ has quit IRC | 14:20 | |
| *** mriedem has joined #openstack-cinder | 14:24 | |
| openstackgerrit | Mate Lakat proposed openstack/cinder master: Create indexes for foreign keys https://review.openstack.org/437313 | 14:26 |
| *** ociuhandu has joined #openstack-cinder | 14:29 | |
| *** lamt has quit IRC | 14:32 | |
| *** Jack_Iv has quit IRC | 14:37 | |
| *** Jack_Iv has joined #openstack-cinder | 14:37 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - volumes part of CG not managed correctly https://review.openstack.org/441199 | 14:38 |
| *** Jack_Iv has quit IRC | 14:41 | |
| *** liliars has joined #openstack-cinder | 14:42 | |
| *** Jack_Iv has joined #openstack-cinder | 14:46 | |
| *** cknight has joined #openstack-cinder | 14:47 | |
| *** mtanino has joined #openstack-cinder | 14:49 | |
| *** Yogi1 has joined #openstack-cinder | 14:49 | |
| *** mlakat_ has joined #openstack-cinder | 14:52 | |
| *** arne_r has joined #openstack-cinder | 14:52 | |
| mlakat_ | Hi cinder experts, I'd like to ask about a change where I introduce some indexes to the database. | 14:52 |
| mlakat_ | What is the policy - do we add indexes to models.py, or just leave them in migrations? | 14:53 |
| mlakat_ | This is the change in question: https://review.openstack.org/#/c/437313/ | 14:53 |
| mlakat_ | It was -1 'd because I added the indexes in models.py (which I believe is the correct approach) - but for the latest patchset I removed it, and tbh I don't quite see how can that work later. | 14:54 |
| mlakat_ | So I would need some expert opinion on it. | 14:54 |
| *** ducttape_ has joined #openstack-cinder | 14:55 | |
| *** sandanar has joined #openstack-cinder | 14:59 | |
| *** ducttape_ has quit IRC | 14:59 | |
| xinli | erlon: @smcginnis: good news that NetApp also fixed CI issue and voted on patch 336092. I also verified that no code conflict today morning. | 15:02 |
| *** nicolasbock has quit IRC | 15:03 | |
| *** lamt has joined #openstack-cinder | 15:03 | |
| *** Julien-zte has joined #openstack-cinder | 15:04 | |
| *** pcaruana has quit IRC | 15:09 | |
| *** lucasxu has joined #openstack-cinder | 15:11 | |
| *** arne_r has quit IRC | 15:14 | |
| *** Jack_Iv has quit IRC | 15:22 | |
| *** ducttape_ has joined #openstack-cinder | 15:22 | |
| *** pcaruana has joined #openstack-cinder | 15:23 | |
| *** Jack_Iv has joined #openstack-cinder | 15:23 | |
| *** dgonzalez has joined #openstack-cinder | 15:24 | |
| *** tomh__ has quit IRC | 15:24 | |
| *** salv-orlando has joined #openstack-cinder | 15:24 | |
| *** isaacb has quit IRC | 15:26 | |
| openstackgerrit | Helen Walsh proposed openstack/cinder master: VMAX driver - Change vendor_name from EMC to Dell EMC https://review.openstack.org/442560 | 15:27 |
| *** Jack_Iv has quit IRC | 15:27 | |
| *** sandanar has quit IRC | 15:33 | |
| *** tommylikehu_ has quit IRC | 15:33 | |
| jungleboyj | mlakat_: Good question. My initial reaction is to agree that we don't touch models.py . | 15:34 |
| jungleboyj | smcginnis: Thoughts ^^ | 15:34 |
| mlakat_ | jungleboyj, but then how do we merge the migrations? is it done by hand? | 15:34 |
| jungleboyj | mlakat_: What do you mean? | 15:35 |
| mlakat_ | https://github.com/openstack/cinder/blob/master/cinder/db/sqlalchemy/migrate_repo/versions/073_cinder_init.py | 15:35 |
| mlakat_ | jungleboyj, what I meant (maybe I'm missing something) is that from time to time we "squash" migrations to one single init script, as above | 15:36 |
| jungleboyj | mlakat_: Correct. | 15:36 |
| jungleboyj | We do that at the end of each release. | 15:37 |
| mlakat_ | So I was thinking how that is created - I though that the script is created by reflecting the schema? | 15:37 |
| mlakat_ | And if it is, then we need to add the indexes to the schema as well (by schema i mean models.py ) | 15:37 |
| mlakat_ | jungleboyj, how do we create the file above! | 15:37 |
| mlakat_ | s/!/? | 15:37 |
| jungleboyj | It is just created by hand. smcginnis takes any new migrations that were created during the release and puts them in the new init file. | 15:38 |
| jungleboyj | So, at the end of Pike, we would take your change and add it to the latest cinder_init migration. | 15:38 |
| mlakat_ | Ah, right. In that case it could work. I thought that it is somehow created from models.py | 15:39 |
| *** alonma has quit IRC | 15:39 | |
| jungleboyj | mlakat_: No, it is not. :-) | 15:39 |
| jungleboyj | mlakat_: It isn't that fancy. | 15:39 |
| *** ducttape_ has quit IRC | 15:39 | |
| *** alonma has joined #openstack-cinder | 15:39 | |
| mlakat_ | I thought that we are aiming for metadata.create_all(engine) == latest migrated database | 15:40 |
| *** ducttape_ has joined #openstack-cinder | 15:40 | |
| mlakat_ | but thinking about that, it would be kind of difficult to achieve given the diverse database implementations. | 15:41 |
| jungleboyj | So, in the future, if we have migrations that add Foreign Keys we should make sure that they also add an index? | 15:41 |
| mlakat_ | jungleboyj, for postgresql, yes. | 15:41 |
| *** ducttape_ has quit IRC | 15:42 | |
| mlakat_ | mysql automatically adds indexes to those, but postgres doesn't | 15:42 |
| *** ducttape_ has joined #openstack-cinder | 15:42 | |
| mlakat_ | jungleboyj, another thing we might want to add is indexes on project_id. | 15:42 |
| mlakat_ | But that's another step. | 15:42 |
| mlakat_ | jungleboyj, I also added this models.py question to tomorrow's meeting, do you think we still need to ask someone else about this, or are we confident that it's the way (I think I am convinced :-) | 15:43 |
| *** psachin has quit IRC | 15:43 | |
| *** alonma has quit IRC | 15:44 | |
| jungleboyj | Yeah, so looking at the code again here is what I think will happen ... | 15:44 |
| jungleboyj | This checks to see if there is an index, if there isn't one, it creates it. | 15:45 |
| mlakat_ | true | 15:45 |
| jungleboyj | At the end of Pike that would get added to our init file so this will always be done in the future. | 15:45 |
| *** eharney has joined #openstack-cinder | 15:45 | |
| *** alonma has joined #openstack-cinder | 15:45 | |
| mlakat_ | right - makes sense. | 15:45 |
| *** alonma has quit IRC | 15:49 | |
| *** crose has joined #openstack-cinder | 15:50 | |
| *** lamt has quit IRC | 15:50 | |
| hemna | mornin | 15:50 |
| openstackgerrit | Justin A Wilson proposed openstack/python-cinderclient master: Fix discover_version https://review.openstack.org/385629 | 15:51 |
| *** alonma has joined #openstack-cinder | 15:52 | |
| *** nkrinner is now known as nkrinner_afk | 15:53 | |
| jungleboyj | hemna: Morning. | 15:53 |
| jungleboyj | mlakat_: Sorry, had a phone call and got distracted. | 15:53 |
| jungleboyj | So, as core reviewers we need to then be aware that any migrations that add indexes in the future should do something similar. | 15:54 |
| hemna | jungleboyj, url? | 15:54 |
| *** diablo_rojo has joined #openstack-cinder | 15:54 | |
| jungleboyj | hemna: https://review.openstack.org/#/c/437313/ | 15:55 |
| jungleboyj | mlakat_: The one thing, though, is we may then need to put the ensure_indexed function somewhere that others can share it. | 15:56 |
| *** alonma has quit IRC | 15:56 | |
| *** chhavi has quit IRC | 15:56 | |
| hemna | shouldn't this be done at the end of cinder manage db sync ? | 15:57 |
| *** chhavi has joined #openstack-cinder | 15:57 | |
| hemna | in addition to the db migration | 15:58 |
| hemna | then we'll never have to worry about it again | 15:58 |
| jungleboyj | hemna: Good thought. | 15:58 |
| jungleboyj | mlakat_: Is there a way to just automatically find all indexes? | 15:58 |
| *** lamt has joined #openstack-cinder | 16:00 | |
| *** alonmarx has joined #openstack-cinder | 16:00 | |
| mlakat_ | jungleboyj, I have some scripts, yes | 16:01 |
| mlakat_ | So what I did: created mysql database -> listed all the indexes | 16:01 |
| mlakat_ | created postgresql database -> listed all the indexes | 16:01 |
| mlakat_ | and compared the two. | 16:01 |
| mlakat_ | jungleboyj, don't you want to automatically find foreign keys instead? | 16:02 |
| mlakat_ | I think I did that as well initially: discover foreign keys, and add indexes where they exist. | 16:03 |
| mlakat_ | Assuming that no multi-column index covers that. | 16:03 |
| hemna | that would be best to do that at the end of cinder-manage db sync | 16:03 |
| hemna | then we wouldn't ever have to worry about it | 16:03 |
| *** lpetrut has quit IRC | 16:03 | |
| mlakat_ | hemna, the other thing we need to find out is if sqlalchemy does it concurrently or not - or is there a policy that db migrations can cause downtime? | 16:04 |
| jungleboyj | mlakat_: I am sorry, yes, I meant foreign keys. | 16:04 |
| jungleboyj | Yeah, if there is a way to do it automatically at migration time, that is the best approach. | 16:05 |
| mlakat_ | by concurrently I meant CREATE INDEX CONCURRENTLY (https://www.postgresql.org/docs/9.1/static/sql-createindex.html) | 16:05 |
| *** alonma has joined #openstack-cinder | 16:06 | |
| mlakat_ | jungleboyj, that would hurt the principle of migrations, being well-defined versions of database schema ? | 16:06 |
| mlakat_ | what I mean is that if after each db sync, we ran an add indexes - that kind of smells, or is it just me? | 16:07 |
| jungleboyj | mlakat_: Does it though? We are just making sure that all foreign keys have indexes. That doesn't really break the schema. | 16:07 |
| *** ducttape_ has quit IRC | 16:07 | |
| *** dave-mcc_ has joined #openstack-cinder | 16:08 | |
| mlakat_ | ok, is there a hook that is executed on each dbsync? do we already have some sort of mechanism for this somewhere? | 16:09 |
| *** ducttape_ has joined #openstack-cinder | 16:09 | |
| mlakat_ | jungleboyj, I need to run now, leaving the item on tomorrow's agenda, that will not hurt. | 16:10 |
| *** pdeore has quit IRC | 16:10 | |
| *** alonma has quit IRC | 16:10 | |
| *** chenhb has quit IRC | 16:10 | |
| jungleboyj | mlakat_: Ok. Yeah, it will be good to discuss there. | 16:10 |
| *** dave-mccowan has quit IRC | 16:10 | |
| *** chenhb has joined #openstack-cinder | 16:10 | |
| *** mlakat_ is now known as mlakat | 16:11 | |
| *** dave-mccowan has joined #openstack-cinder | 16:11 | |
| *** e0ne has quit IRC | 16:12 | |
| *** dave-mcc_ has quit IRC | 16:12 | |
| *** akapil_ has joined #openstack-cinder | 16:17 | |
| *** akapil_ has quit IRC | 16:17 | |
| *** redondo-mk has left #openstack-cinder | 16:20 | |
| *** akapil has quit IRC | 16:20 | |
| *** redondo-mk has joined #openstack-cinder | 16:20 | |
| *** belmoreira has quit IRC | 16:22 | |
| *** crushil has quit IRC | 16:22 | |
| *** Julien-zte has quit IRC | 16:25 | |
| *** Julien-zte has joined #openstack-cinder | 16:25 | |
| *** Julien-zte has quit IRC | 16:25 | |
| *** Julien-zte has joined #openstack-cinder | 16:26 | |
| *** pcaruana has quit IRC | 16:26 | |
| *** Julien-zte has quit IRC | 16:26 | |
| *** Julien-zte has joined #openstack-cinder | 16:27 | |
| *** Julien-zte has quit IRC | 16:28 | |
| *** Julien-zte has joined #openstack-cinder | 16:28 | |
| *** Julien-zte has quit IRC | 16:29 | |
| *** esker has joined #openstack-cinder | 16:29 | |
| *** esker has quit IRC | 16:29 | |
| *** Julien-zte has joined #openstack-cinder | 16:29 | |
| *** Julien-zte has quit IRC | 16:29 | |
| *** rcernin has quit IRC | 16:30 | |
| *** chhavi_ has joined #openstack-cinder | 16:33 | |
| *** chhavi has quit IRC | 16:35 | |
| *** eharney_ has joined #openstack-cinder | 16:36 | |
| *** markvoelker has quit IRC | 16:37 | |
| *** lucasxu has quit IRC | 16:38 | |
| *** tesseract has quit IRC | 16:38 | |
| *** eharney has quit IRC | 16:38 | |
| *** tushar_ has quit IRC | 16:38 | |
| redondo-mk | Hi. I had issue with EMC Unity because of older version that is not supported (4.0.1) but now after the upgrade to 4.1.1 things are still not working as expected. Creating a volume from image is not stable and creating a volume from a volume does not work (error in dd command execution). | 16:38 |
| *** eharney_ is now known as eharney | 16:39 | |
| redondo-mk | I'm referring to this: https://bugs.launchpad.net/cinder/+bug/1662584 | 16:39 |
| openstack | Launchpad bug 1662584 in Cinder "EMC Unity Volume Cloning " [High,Fix released] - Assigned to Ryan Liang (ryan-liang) | 16:39 |
| *** JoseMello has joined #openstack-cinder | 16:39 | |
| xinli | xyang1: thank you very much for fixng the EMC CI broken issue and EMC CI has voted on patch 336092. Could you please help to review the patch: https://review.openstack.org/#/c/336092/ | 16:39 |
| *** lucasxu has joined #openstack-cinder | 16:40 | |
| xyang1 | xinli: sure | 16:40 |
| redondo-mk | I would guess that the issue is not in the driver or cinder but I'm not sure... | 16:40 |
| xinli | xyang1: thanks, Most of 3rd party CIs broken issues have been resolved and the CIs have voted on this patch. I will continue follow up HPMSA, HPE, Nexenta, Pure(before the CI broken occurred, patch 336092 passed all the listed CIs tests) | 16:41 |
| xyang1 | xinli: ok, thanks | 16:41 |
| *** alonma has joined #openstack-cinder | 16:42 | |
| *** alonma has quit IRC | 16:46 | |
| *** crushil has joined #openstack-cinder | 16:48 | |
| *** nicolasbock has joined #openstack-cinder | 16:52 | |
| *** alonma has joined #openstack-cinder | 16:53 | |
| *** strigazi is now known as strigazi_AFK | 16:53 | |
| *** isaacb has joined #openstack-cinder | 16:54 | |
| *** alonma has quit IRC | 16:58 | |
| *** aarefiev is now known as aarefiev_afk | 17:00 | |
| *** chhavi_ has quit IRC | 17:02 | |
| openstackgerrit | Merged openstack/os-brick master: Add Python 3.5 classifier and venv https://review.openstack.org/423032 | 17:04 |
| *** alonma has joined #openstack-cinder | 17:04 | |
| *** alonma has quit IRC | 17:09 | |
| *** savihou has quit IRC | 17:09 | |
| *** markvoelker has joined #openstack-cinder | 17:12 | |
| *** ntpttr has joined #openstack-cinder | 17:15 | |
| geguileo | jungleboyj: patrickeast hemna Could you have another look at the multipath friendly names patch? https://review.openstack.org/433102 | 17:16 |
| *** lucasxu has quit IRC | 17:16 | |
| *** ntpttr_ has joined #openstack-cinder | 17:16 | |
| bluex | do we still support config file based default quota limits, or can I remove code that support this (deprecated since 2013) | 17:17 |
| jungleboyj | geguileo: Sure looking. | 17:20 |
| geguileo | jungleboyj: thanks! | 17:20 |
| *** isaacb has quit IRC | 17:25 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/os-brick master: Updated from global requirements https://review.openstack.org/432055 | 17:25 |
| *** Jack_V has joined #openstack-cinder | 17:25 | |
| *** Apoorva has joined #openstack-cinder | 17:26 | |
| *** Jack_V has quit IRC | 17:28 | |
| *** Jack_I has quit IRC | 17:29 | |
| *** Jack_I has joined #openstack-cinder | 17:29 | |
| *** nikeshm has joined #openstack-cinder | 17:30 | |
| *** lucasxu has joined #openstack-cinder | 17:31 | |
| *** prashkre has joined #openstack-cinder | 17:38 | |
| *** ntpttr_ has quit IRC | 17:39 | |
| *** erhudy has joined #openstack-cinder | 17:42 | |
| nikeshm | hi | 17:46 |
| nikeshm | hi in my CI env, quota is not updating sometimes | 17:46 |
| nikeshm | https://ask.openstack.org/en/question/494/how-to-reset-incorrect-quota-count/ | 17:46 |
| nikeshm | did any one face the issue | 17:46 |
| nikeshm | is there any good solution rather than logging in nova database and do some reset | 17:47 |
| *** Jack_Iv has joined #openstack-cinder | 17:47 | |
| nikeshm | patrickeast: did you face the issue | 17:48 |
| *** sudipto has joined #openstack-cinder | 17:48 | |
| *** sudipto_ has joined #openstack-cinder | 17:48 | |
| nikeshm | patrickeast: BTW, i found and solved the problem due to which our CI was not running cherry picking the cinder patch code automatically | 17:49 |
| nikeshm | some dns issue | 17:49 |
| *** Jack_Iv has quit IRC | 17:50 | |
| *** Jack_Iv has joined #openstack-cinder | 17:51 | |
| *** david-lyle_ has joined #openstack-cinder | 17:52 | |
| nikeshm | smcginnis patrickeast: after correcting dns issue inside nodepool VM, Kaminario CI is running tests on top of the patch code which is under investigation, now it is working as expected :) | 17:54 |
| *** david-lyle has quit IRC | 17:54 | |
| *** david-lyle_ is now known as david-lyle | 17:55 | |
| *** Jack_Iv has quit IRC | 17:57 | |
| *** ducttape_ has quit IRC | 17:59 | |
| *** sudipto has quit IRC | 18:00 | |
| *** sudipto_ has quit IRC | 18:00 | |
| jgriffith | Hey hemna what's the link to that fancy whiz bang tool you have to check 3'rd party CI stats? | 18:01 |
| *** links has joined #openstack-cinder | 18:01 | |
| hemna | jgriffith, https://github.com/openstack/third-party-ci-tools/blob/master/monitoring/lastcomment-scoreboard/cireporter.py | 18:01 |
| jgriffith | hemna thanks! | 18:01 |
| hemna | np | 18:02 |
| smcginnis | jgriffith: Regularly updated output, if that's all you're looking for: http://cinderstats-dellstorage.rhcloud.com/cireport.txt | 18:02 |
| hemna | there is also this: | 18:02 |
| hemna | https://github.com/openstack/third-party-ci-tools/blob/master/monitoring/lastcomment-scoreboard/cinder_ci.py | 18:02 |
| *** dhinesh has joined #openstack-cinder | 18:02 | |
| hemna | that takes the input of cinder's generate_driver_list | 18:02 |
| hemna | and runs all the CI's listed in there. | 18:02 |
| hemna | or the cireporter.py can run against 1 or N Ci's | 18:03 |
| smcginnis | hemna: Hmm, I'm running ci_reporter. Is cinder_ci basically the same, but dynamically updated? | 18:04 |
| *** david-lyle has quit IRC | 18:04 | |
| *** ducttape_ has joined #openstack-cinder | 18:04 | |
| *** david-lyle has joined #openstack-cinder | 18:05 | |
| *** _cjones_ has joined #openstack-cinder | 18:06 | |
| jgriffith | I was running the reporter just to isolate mine for now | 18:13 |
| *** chhavi_ has joined #openstack-cinder | 18:13 | |
| jgriffith | 96% average out of the last 50 W00T | 18:13 |
| jgriffith | 82% out of the last 200, not terrible | 18:13 |
| jgriffith | and I'm even posting the logs again now | 18:14 |
| jgriffith | so I can go back to ignoring it again until the first week of the month :) | 18:14 |
| smcginnis | jgriffith: I'm guessing that's better than average, | 18:14 |
| jgriffith | @smcginnis yeah, I haven't looked... just have a recurring thing on my calendar to "inspect CI" the first Monday of each month" :) | 18:15 |
| *** _cjones_ has quit IRC | 18:15 | |
| smcginnis | hah! | 18:15 |
| *** _cjones_ has joined #openstack-cinder | 18:20 | |
| *** david-lyle has quit IRC | 18:23 | |
| *** david-lyle has joined #openstack-cinder | 18:25 | |
| jungleboyj | jgriffith: Nice! Can you teach other CI owners? ;-) | 18:27 |
| *** david-lyle has quit IRC | 18:29 | |
| jgriffith | jungleboyj sure.... https://support.google.com/calendar/answer/37115?co=GENIE.Platform%3DDesktop&hl=en | 18:30 |
| jgriffith | not hard :) | 18:30 |
| hemna | smcginnis, cinder_ci consumes the JSON output of cinder's generate_driver_list.py > cinder_drivers.json | 18:31 |
| hemna | and uses that data to determine which CI's to report on with cireporter.py | 18:31 |
| hemna | so run generate_driver_list.py first, output that to a file. use that file to feed into cinder_ci.py | 18:32 |
| hemna | and you'll get whatever cinder drivers there are and their associated CIs (from CI_WIKI_NAME) | 18:33 |
| *** david-lyle has joined #openstack-cinder | 18:36 | |
| *** mvk has quit IRC | 18:36 | |
| jungleboyj | jgriffith: Smart Alec. | 18:37 |
| jgriffith | jungleboyj :) | 18:38 |
| nikeshm | did any one face https://ask.openstack.org/en/question/494/how-to-reset-incorrect-quota-count/ | 18:39 |
| *** salv-orlando has quit IRC | 18:39 | |
| *** eharney has quit IRC | 18:41 | |
| *** salv-orlando has joined #openstack-cinder | 18:44 | |
| *** eharney has joined #openstack-cinder | 18:45 | |
| *** xyang has joined #openstack-cinder | 18:46 | |
| nikeshm | good n8 | 18:50 |
| *** nikeshm has quit IRC | 18:50 | |
| jungleboyj | smcginnis: We can add config options with a backport, we just can remove or change them. Correct? | 18:52 |
| smcginnis | jungleboyj: Good point. I believe so, but let me take another look at the guidelines. | 18:52 |
| jungleboyj | Ok. I am pretty sure that is the case. | 18:52 |
| smcginnis | jungleboyj: Yeah, I think this is the only issue to watch out for, in that respect: "Incompatible config file changes" | 18:53 |
| smcginnis | jungleboyj: https://docs.openstack.org/project-team-guide/stable-branches.html#review-guidelines | 18:54 |
| jungleboyj | smcginnis: That was how I remembered it. Thanks for confirming. :-) | 18:54 |
| smcginnis | jungleboyj: Good to clarify that. ;) | 18:55 |
| *** ducttape_ has quit IRC | 18:56 | |
| jungleboyj | Ok those backports are taken care of. | 18:57 |
| *** adisky_ has quit IRC | 18:59 | |
| *** asselin has joined #openstack-cinder | 18:59 | |
| *** sdague has quit IRC | 19:02 | |
| *** david-lyle has quit IRC | 19:04 | |
| *** david-lyle has joined #openstack-cinder | 19:06 | |
| *** mvk has joined #openstack-cinder | 19:07 | |
| *** crushil has quit IRC | 19:08 | |
| *** alonma has joined #openstack-cinder | 19:11 | |
| *** ducttape_ has joined #openstack-cinder | 19:13 | |
| *** _ducttape_ has joined #openstack-cinder | 19:14 | |
| *** alonma has quit IRC | 19:15 | |
| *** ducttape_ has quit IRC | 19:18 | |
| *** crushil has joined #openstack-cinder | 19:18 | |
| *** _ducttape_ has quit IRC | 19:18 | |
| *** markstur has joined #openstack-cinder | 19:19 | |
| *** xyang has quit IRC | 19:19 | |
| *** salv-orl_ has joined #openstack-cinder | 19:22 | |
| erlon | xinli: hey, have reviewed the patch. Thanks for the work on it. Its almost good, just need to address the comments I added. | 19:24 |
| *** salv-orlando has quit IRC | 19:25 | |
| *** Jack_Iv has joined #openstack-cinder | 19:26 | |
| erlon | jgriffith: can you also check this (https://review.openstack.org/#/c/336092/) when you have a chance? You suggested this approach at first so, good to have your opinion. | 19:27 |
| prashkre | Hi. Could you please someone check the bug created https://bugs.launchpad.net/cinder/+bug/1670801. | 19:33 |
| openstack | Launchpad bug 1670801 in Cinder "Filters are not applied on Consistency group snapshots GET API" [Undecided,New] | 19:33 |
| prashkre | is it valid or due to any constraints the filters are not applied? | 19:34 |
| xyang1 | prashkre: they are not implemented. There is a patch to add it to group snapshots though | 19:35 |
| xyang1 | prashkre: https://review.openstack.org/#/c/441651/ | 19:37 |
| openstackgerrit | christopher.uhler proposed openstack/os-brick master: os-brick connector for Veritas HyperScale https://review.openstack.org/442754 | 19:39 |
| *** ChubYann has joined #openstack-cinder | 19:43 | |
| *** xyang has joined #openstack-cinder | 19:44 | |
| prashkre | xyang1: Thanks for pointing me to https://review.openstack.org/#/c/441651/. | 19:45 |
| *** xyang has quit IRC | 19:45 | |
| *** redondo-mk has quit IRC | 19:45 | |
| prashkre | xyang1: so this is available for v3 when patch merges. I guess we need to support the v2 API as well? | 19:47 |
| *** xyang has joined #openstack-cinder | 19:47 | |
| xyang1 | prashkre: group snapshots are only introduced in v3. Also we are talking about deprecating CG API some day, so I don't think we need to add this there | 19:49 |
| prashkre | xyang1: Thanks for info. when can we expect this patch will merge? | 19:52 |
| *** crose has quit IRC | 19:52 | |
| xyang1 | prashkre: this was just submitted. Will go thru the normal code review process. Can't tell you exactly when it will be merged, but should be available in pike | 19:53 |
| prashkre | xyang1: sure. Thank you!. | 19:55 |
| xyang1 | prashkre: you are welcome | 19:55 |
| *** e0ne has joined #openstack-cinder | 19:57 | |
| *** lucasxu has quit IRC | 19:57 | |
| *** chhavi_ has quit IRC | 20:00 | |
| *** gouthamr has quit IRC | 20:01 | |
| *** salv-orl_ has quit IRC | 20:03 | |
| Swanson | xyang1, How do I tell if a group is of a cg type? | 20:04 |
| Swanson | xyang1, or that has a consistent snapshot property? | 20:04 |
| *** lucasxu has joined #openstack-cinder | 20:04 | |
| xyang | Swanson: are you using Groups command or CG command? Is this a migrated group from CG? | 20:04 |
| Swanson | xyang, I'm coding create_group() | 20:05 |
| xyang | Swanson: your group type should have consistent_group_snapshot_enabled set to "<is> True" | 20:06 |
| Swanson | xyang, you're way more useful than smcginnis. Thanks! | 20:07 |
| xyang | Swanson: :) | 20:07 |
| smcginnis | I'm good with that. | 20:07 |
| xyang | Swanson: Take a look of this section: https://github.com/openstack/cinder/blob/master/doc/source/devref/groups.rst#group-type-and-group-specs--volume-types-and-extra-specs | 20:07 |
| *** ducttape_ has joined #openstack-cinder | 20:07 | |
| *** ducttape_ has quit IRC | 20:08 | |
| *** ducttape_ has joined #openstack-cinder | 20:08 | |
| *** links has quit IRC | 20:09 | |
| *** wN has joined #openstack-cinder | 20:12 | |
| *** wN has quit IRC | 20:12 | |
| *** wN has joined #openstack-cinder | 20:12 | |
| *** simondodsley has joined #openstack-cinder | 20:13 | |
| *** dustins has quit IRC | 20:15 | |
| jgriffith | erlon still around? | 20:15 |
| jgriffith | erlon I don't really remember the conversation on https://review.openstack.org/#/c/336092/29 | 20:15 |
| jgriffith | erlon I'm not sure I follow exactly what you have going on there, or if it's the best option | 20:16 |
| jgriffith | erlon I do sort of recall something more along the lines of checking the volume-size when the existing create call returns, and if it's not correct issue the extend call | 20:16 |
| erlon | jgriffith: that was about creating a commom method to drivers to fix the create_from_* problem | 20:16 |
| jgriffith | erlon right... but I think what I had in mind was a bit different | 20:17 |
| jgriffith | and didn't change the method or the drivers | 20:17 |
| erlon | jgriffith: initially I implemented in flow_manager, than you suggested to add a common function in the base class | 20:17 |
| jgriffith | erlon sorry... I honestly don't remember the details of what we talked about there, and I guess the problem is that the drivers are broke so my statement about checking and extending doesn't actually work either | 20:18 |
| erlon | jgriffith: yeah the thing is, several drivers solved the bug in their drivers code, while another bunch no, so, we either would fix driver by driver or implement a common way, and 'revert' the fixes drivers had added. Thats why we neeed to change so many drivers | 20:19 |
| jgriffith | erlon so I don't follow exactly how your patch works I'm afraid | 20:19 |
| jgriffith | erlon well, what sucks is that we 'don't know' if a driver did the right thing or not | 20:19 |
| jgriffith | Unless we create a tempest test that attaches the volume and checks the size from the Instance to make sure it matches what Cinder reports | 20:20 |
| Swanson | xyang, create_group_from_src now required? | 20:20 |
| xyang | Swanson: yes | 20:20 |
| jgriffith | erlon anyway... back to trying to fix it in the flow etc... | 20:20 |
| erlon | jgriffith: so, to recap, the problem happens when a driver tries to create a volume from a snapshot or other volume that have a a smaller size than the volume being created | 20:20 |
| erlon | jgriffith: this is anoother problem | 20:20 |
| jgriffith | I'm not sure how you're code does this exactly other than you inspected the driver yourself and tried to interpret how it works | 20:20 |
| erlon | jgriffith: checking if the backend is really extending | 20:21 |
| jgriffith | erlon oh, I'm familiar with the problem, I'm having difficulty getting my head wrapped around the proposed fix though :) | 20:21 |
| erlon | jgriffith: yes, tough work, but that was the approach | 20:21 |
| jgriffith | erlon right, but now we have kind of an ugly thing... | 20:21 |
| jgriffith | erlon we have this function variable thing you set up based on drivers that did the right thing, and others use the old method still | 20:22 |
| *** crose has joined #openstack-cinder | 20:22 | |
| erlon | jgriffith: we implemented a common behavior in the base class. The driver creates a volume and if needed it extend right after, by calling driver.extend() | 20:22 |
| jgriffith | erlon what I mean is, it *might* fix the actual problem, but in turn it kind of creates a new one | 20:22 |
| jgriffith | erlon but only *some* of the drivers use that method near as I can tell.... others still use the old one | 20:23 |
| openstackgerrit | Eric Harney proposed openstack/cinder master: Tests: Use six.moves.urllib urlencode https://review.openstack.org/442769 | 20:23 |
| jgriffith | and the "calling extend if needed" is the part that I'm saying other than looking at the driver code I'm not sure now how you determine that? | 20:23 |
| erlon | jgriffith: looking at the driver code and checking if the CI is passing after the fix | 20:23 |
| jgriffith | erlon I dislike having unique paths based on the driver being called is my point I guess | 20:24 |
| jgriffith | erlon so if you wanted to do something like what you have here I'd like to offer an alternative suggestion if you don't mind... | 20:24 |
| erlon | jgriffith: thats why I firstly checked all CI results, and now xinli did the same | 20:24 |
| openstackgerrit | Ken'ichi Ohmichi proposed openstack/cinder master: Switch to use stable data_utils https://review.openstack.org/442770 | 20:25 |
| jgriffith | erlon implement a new driver method "verify_volume_size" | 20:25 |
| erlon | jgriffith: what you mean? | 20:25 |
| erlon | jgriffith: with the unique paths | 20:25 |
| jgriffith | erlon https://review.openstack.org/#/c/336092/29/cinder/volume/driver.py | 20:26 |
| jgriffith | Line #574 and 581 | 20:26 |
| *** _cjones_ has quit IRC | 20:27 | |
| jgriffith | oh wait... no, that's not it | 20:27 |
| jgriffith | just a sec | 20:27 |
| erlon | jgriffith: hmm, got it | 20:27 |
| jgriffith | erlon so like this is confusing to me. https://review.openstack.org/#/c/336092/29/cinder/volume/drivers/lvm.py | 20:28 |
| jgriffith | erlon so that base driver does exactly what it's supposed to... but now we break it out into 2 seperate driver calls using a clone_func variable | 20:29 |
| *** gouthamr has joined #openstack-cinder | 20:29 | |
| erlon | jgriffith: honestly I dind't liked the final result as well, having to touch all drivers that already had the problem solved. That would be more acceptable if was done before everyone had its own solution, but, I think xinli fall in this patch I did because I believe he was trying to fix the easy way and someone pointed to a common solution again | 20:30 |
| jgriffith | I also don't quite understand, the SolidFire driver for example is another one that does things correctly; but it's not changed here | 20:30 |
| jgriffith | erlon yeah, sadly I can't tell you how many times I have what I think is going to be a great and elegant solution, but once I get done with it I'm not as happy :( | 20:30 |
| jgriffith | erlon Yeah, fixing in the driver without actually modifying the driver might not be something we can figure out now that I think about it more | 20:31 |
| openstackgerrit | Merged openstack/os-brick master: Fix multipath flush when using friendly names https://review.openstack.org/433102 | 20:31 |
| jgriffith | and if we have to go to that level then the driver should just be fixed :) | 20:31 |
| jgriffith | in other words, as much time and effort as you put into this inspecting and working each driver, maybe it would be just as easy to fix the 100 drivers by adding the check and extend call inside of them? | 20:32 |
| jgriffith | Then making sure we have a tempest test that exposes the failure | 20:32 |
| jgriffith | I dunno | 20:32 |
| erlon | jgriffith: you raised a good point in the LVM driver, it seems it should only have the extend removed, and not add the function to overhide the default behaviour | 20:32 |
| jgriffith | the problem with so many drivers is there's no easy way to fix this or advance the ball | 20:32 |
| jgriffith | we've kinda screwed ourselves :) | 20:33 |
| erlon | jgriffith: man I kill you!! | 20:34 |
| *** Jack_Iv has quit IRC | 20:34 | |
| *** Jack_Iv has joined #openstack-cinder | 20:34 | |
| erlon | jgriffith: haha | 20:34 |
| *** Jack_Iv has quit IRC | 20:35 | |
| *** JoseMello has quit IRC | 20:35 | |
| erlon | jgriffith: well, in the end, drivers that didnt fixed that already are the one that are not actively maintaining | 20:35 |
| smcginnis | The problem is, some drivers can more optimally get the new, correct, size during the clone. | 20:35 |
| *** Jack_Iv has joined #openstack-cinder | 20:35 | |
| smcginnis | And others have to do the fallback of just extending after clonging. | 20:35 |
| smcginnis | I think when we discussed this in the past, really the only way to do the right thing is to change the driver interface. | 20:36 |
| smcginnis | And add the smarts into that interaction so the driver can say whether it was able to get the new size or not. | 20:36 |
| erlon | smcginnis: I know that, and the patch handles those cases too | 20:36 |
| openstackgerrit | Merged openstack/os-brick master: Retry multipath flush when map is in use https://review.openstack.org/433103 | 20:37 |
| erlon | smcginnis: the problem here is, what is better, to remove snipped on drivers that already individually fixed the problem and then use the commom solution, or just add the extend function to the others that havent fixed that already? | 20:39 |
| erlon | smcginnis: if this was solution was put when we first reported the problem, and no drivers had that fixed, the commom method would me more appealing, now, is not that much anymore | 20:40 |
| smcginnis | erlon: Well, my point isn't about drivers that have "fixed the problem". It's about ones that can more optimally create the larger cloned volume and would therefore be more inefficient by resorting to a generic approach. | 20:40 |
| *** ducttape_ has quit IRC | 20:40 | |
| *** Jack_Iv has quit IRC | 20:40 | |
| *** Jack_Iv has joined #openstack-cinder | 20:40 | |
| *** rhe00 has quit IRC | 20:41 | |
| erlon | smcginnis: hmm, so, those ones are the ones that created the volumes using its optimized way with the correct size. For those we skip the generic way: https://review.openstack.org/#/c/336092/29/cinder/volume/drivers/violin/v7000_fcp.py | 20:42 |
| erlon | smcginnis: jgriffith : xinli : got to go, we pick that later | 20:43 |
| smcginnis | erlon: But as it is now, like jgriffith pointed out, that requires knowledge of the driver and whether they can do that or not. | 20:43 |
| *** Jack_Iv_ has joined #openstack-cinder | 20:43 | |
| *** Jack_Iv has quit IRC | 20:47 | |
| xinli | @smcginnis: so far only two vendors have really fixed the issues 1) IBM Storawize 2) VMAX, all the other dirvers have not fixed the issue. | 20:49 |
| smcginnis | xinli: The "issue" being, being able to create a clone that is larger than the original? | 20:54 |
| xinli | @smcginnis: yes, regarding the 336092 | 20:55 |
| *** prashkre has quit IRC | 20:55 | |
| *** dustins has joined #openstack-cinder | 20:55 | |
| smcginnis | xinli: That is not correct. Quite a few drivers actually handle that correctly. | 20:56 |
| xinli | @smcginnis: in this case, what is the next step? just leaving the drivers which handles correctly ? | 20:57 |
| smcginnis | xinli: Well, I'm not sure really. | 20:58 |
| smcginnis | xinli: I think to really fix this in an optimal way would require changing the driver interface. | 20:58 |
| jgriffith | xinli erlon @smcginnis I think we have a couple options: | 20:58 |
| jgriffith | to @smcginnis we could change the interface | 20:58 |
| jgriffith | or | 20:58 |
| xinli | @smcginnis: at least run all the driver code aginst CI and make all the CIs issue resolved. : ) | 20:59 |
| jgriffith | just go on a rampage and fix the drivers manually | 20:59 |
| jgriffith | or | 20:59 |
| *** xyang has quit IRC | 20:59 | |
| jgriffith | implement a verification check in tempest, expose the losers and ridicule them mercilessly | 20:59 |
| jgriffith | @erlon I'm fine with what you were doing if we just do as @smcginnis pointed out and make it a *new* interface and remove the old one, but the problem I have with that is the point about some doing it when they actually issue the clone call etc. So it gets clunky no matter what. IMHO I think the driver should just be responsible to do what it's supposed to. and in the future we should be better when | 21:01 |
| jgriffith | reviewing to make sure this sort of thing doesn't slip through the cracks (or again.. a verification test for it) | 21:01 |
| *** Jack_I has quit IRC | 21:02 | |
| jgriffith | erlon I kinda feel crappy because you obviously put a ton of work into that patch | 21:02 |
| smcginnis | jgriffith: I actually kind of like the idea of adding a scenario test that does what you described. | 21:02 |
| *** sdague has joined #openstack-cinder | 21:03 | |
| smcginnis | jgriffith: Create a volume, write to it, detach, clone it to a new volume of a larger size, attach, verify new size. | 21:03 |
| smcginnis | Something like that. | 21:03 |
| jgriffith | @smcginnis @erlon there's one other option that we haven't mentioned that fixes ALL of this | 21:03 |
| smcginnis | jgriffith: Just use LVM? | 21:03 |
| jgriffith | haah! Well that's not bad.. but no :) | 21:03 |
| jgriffith | so... | 21:03 |
| jgriffith | just disable clone and create from snapshot of different size in cinder/volume/api.py | 21:04 |
| jgriffith | don't even let it be a thing anymore | 21:04 |
| jgriffith | user can still extend the volume on their own if they like | 21:04 |
| jgriffith | this chaining things together to make the user *happier* sometimes isn't really worth it IMO | 21:04 |
| jgriffith | That would *fix* the problem, and not break anything at the same time | 21:05 |
| jgriffith | well.. not break anything on our side :) | 21:05 |
| smcginnis | Well, kind of hard to disable cloning now. But I think some drivers did have the same issue with create from snapshot. At least at one point. | 21:06 |
| jgriffith | @smcginnis no no.. not saying dsiable cloning | 21:06 |
| jgriffith | disable | 21:06 |
| smcginnis | Just internally make it a create from snapshot? | 21:06 |
| *** lpetrut has joined #openstack-cinder | 21:07 | |
| jgriffith | no... | 21:07 |
| jgriffith | I mean enforce that the size MUST be the same | 21:07 |
| jgriffith | so like "if src_sz != dest_sz: PUKE" | 21:07 |
| jgriffith | I don't remember all the conversations and arguments around not just doing that. I may have even been one to argue in favor of allowing an auto-extend (but I don't think so) | 21:08 |
| jgriffith | users have the API's to do all of these things, we don't necessarily gain much by doing it all for them IMO | 21:09 |
| *** xyang has joined #openstack-cinder | 21:09 | |
| smcginnis | jgriffith: I don't think that would help us now. | 21:09 |
| smcginnis | jgriffith: It would be an API change. | 21:09 |
| smcginnis | jgriffith: So we could add a mv that does it that way. | 21:09 |
| jgriffith | @smcginnis yeah | 21:09 |
| jgriffith | but this shows how useless mv's are | 21:09 |
| smcginnis | jgriffith: But I think we would still need to support the old way. | 21:09 |
| jgriffith | then we just have to support "MORE" things | 21:09 |
| smcginnis | jgriffith: Yeah, just cumulative, at least so far. | 21:10 |
| jgriffith | we're really screwed | 21:10 |
| jgriffith | hub clone openstack/cinder | 21:10 |
| jgriffith | grrr | 21:10 |
| *** alonma has joined #openstack-cinder | 21:12 | |
| xinli | jgriffith: @smcginnis: at this time, we need to hold and I will wait for your final decision. | 21:13 |
| *** eharney has quit IRC | 21:13 | |
| smcginnis | xinli: Sorry for not having a clear answer for you. | 21:14 |
| jgriffith | https://github.com/openstack/cinder/commit/a953ecea5b4ad74ae28c8306b88e01ae6be23740 | 21:15 |
| xinli | @smcginnis: it's OK, I also need to learn the lesson, the good thing is that this patch need to pass all the 3rd CIs, and I have worked with all of them to make most of the CI working now : ) | 21:15 |
| smcginnis | xinli: Yes, thank you for all you did trying to get CIs to pass. That got a lot of maintainers to pay more attention to their broken CIs. | 21:15 |
| jgriffith | https://review.openstack.org/#/c/201754/ | 21:16 |
| smcginnis | xinli: That's very appreciated. | 21:16 |
| *** alonma has quit IRC | 21:17 | |
| jgriffith | I'm not even sure if this was an intentional *feature* :( | 21:17 |
| smcginnis | jgriffith: I don't think it was. | 21:17 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/os-brick master: Updated from global requirements https://review.openstack.org/432055 | 21:17 |
| jgriffith | Well, I'll let the API working group figure out how to deal with stuff like this as per their view of the world | 21:18 |
| jgriffith | seems like the only option now is to fix the broken drivers | 21:19 |
| jgriffith | or fork cinder :) | 21:19 |
| smcginnis | :D | 21:19 |
| *** e0ne has quit IRC | 21:20 | |
| smcginnis | Really, I think there were maybe just a handful of drivers that did not accomodate for the larger destination size now. | 21:20 |
| smcginnis | So a tempest test that would make their CIs fail would probably get their attention pretty quick and get it fixed. | 21:21 |
| smcginnis | Or get them removed. Either way, problem solved. | 21:21 |
| smcginnis | We've been pretty good on new driver reviews of checking for it. | 21:21 |
| smcginnis | So I think the newer drivers should all be good. | 21:21 |
| smcginnis | Tempest tests would make sure they all are correct. | 21:22 |
| *** alonma has joined #openstack-cinder | 21:22 | |
| jgriffith | @smcginnis agreed | 21:25 |
| smcginnis | Looks like there is an API test that gets the volume and verifies the size, but that's just what Cinder has in the DB. | 21:26 |
| smcginnis | We would need a full scenario test that actually attaches and looks at the volume. | 21:26 |
| *** alonma has quit IRC | 21:26 | |
| jgriffith | @smcginnis right | 21:26 |
| jgriffith | @smcginnis I guess I was under the impression that erlon had already identified all of the culprits based on the Bugs in the commit: https://review.openstack.org/#/c/336092/29 | 21:27 |
| jgriffith | or *you* did :) | 21:27 |
| smcginnis | jgriffith: I think most(all?) of those are just ones I filed when we first realized this was how it worked and that we needed to fix some drivers. | 21:28 |
| jgriffith | so 18 | 21:28 |
| jgriffith | https://bugs.launchpad.net/cinder/+bugs?field.tag=create-volume-from-snapshot | 21:28 |
| smcginnis | jgriffith: I went through and filed a bug for every driver that I didn't see explicitly checking the requested size. | 21:28 |
| smcginnis | jgriffith: I do think some of those have already been fixes and/or actually do handle it correctly. | 21:28 |
| jgriffith | @smcginnis that's a pretty admirable effort! | 21:29 |
| *** eharney has joined #openstack-cinder | 21:29 | |
| smcginnis | jgriffith: It was all based on my quick survey of the driver code. | 21:29 |
| jgriffith | oh.. well that's good | 21:29 |
| jgriffith | so let's split it up | 21:29 |
| smcginnis | jgriffith: Sadly, when I filed those I assumed they would all get jumped on and resolved by the end of that cycle. | 21:29 |
| jgriffith | get volunteers tomorrow, split it up based on alphabet | 21:29 |
| smcginnis | but that was kilo or mitaka I think. | 21:29 |
| jgriffith | 25/num_volunteers | 21:29 |
| smcginnis | :) | 21:30 |
| jgriffith | check every driver | 21:30 |
| jgriffith | or take driver-list/num-volunteers | 21:30 |
| jgriffith | create a comittee.. or TigerTeam or whatever you want to call it | 21:30 |
| jgriffith | just fix this crap and move on | 21:30 |
| jgriffith | or not | 21:31 |
| jgriffith | just an idea | 21:31 |
| smcginnis | Oooh, tiger teams. Exciting! :D | 21:32 |
| *** Yogi1 has quit IRC | 21:36 | |
| *** gouthamr has quit IRC | 21:37 | |
| *** salv-orlando has joined #openstack-cinder | 21:38 | |
| *** ducttape_ has joined #openstack-cinder | 21:40 | |
| *** frontrunner has joined #openstack-cinder | 21:41 | |
| *** shaner has quit IRC | 21:41 | |
| *** shaner has joined #openstack-cinder | 21:43 | |
| *** redondo-mk has joined #openstack-cinder | 21:44 | |
| *** ducttape_ has quit IRC | 21:46 | |
| *** asselin has quit IRC | 21:47 | |
| *** crose has quit IRC | 21:48 | |
| *** cknight has quit IRC | 21:51 | |
| *** redondo-mk has quit IRC | 21:52 | |
| *** dave-mcc_ has joined #openstack-cinder | 21:53 | |
| *** xyang has quit IRC | 21:54 | |
| *** dave-mccowan has quit IRC | 21:54 | |
| *** xyang has joined #openstack-cinder | 21:56 | |
| *** gouthamr has joined #openstack-cinder | 21:59 | |
| *** crushil has quit IRC | 22:00 | |
| *** xyang has quit IRC | 22:02 | |
| *** lucasxu has quit IRC | 22:02 | |
| *** lucasxu has joined #openstack-cinder | 22:03 | |
| *** dave-mcc_ has quit IRC | 22:03 | |
| *** porrua_ has quit IRC | 22:06 | |
| xinli | @smcginnis: I also fixed all your asw ell as erlon's comments and try to commit it, is it OK. | 22:11 |
| smcginnis | xinli: Sure, if you want to submit those changes, if we are able to move forward in some way with that it doesn't hurt. | 22:13 |
| *** cknight has joined #openstack-cinder | 22:13 | |
| xinli | @smcginnis: based on your talk with jgriffith, do we need to split the bugs in several groups? | 22:14 |
| *** ducttape_ has joined #openstack-cinder | 22:16 | |
| smcginnis | xinli: If we end up just having each driver make sure they handle the size change correctly, then there would be a patch per bug. | 22:17 |
| smcginnis | xinli: If you're interested, I think the best thing would be to have a tempest scenario test that verifies this. | 22:18 |
| smcginnis | xinli: I think we need to do something like this scenario test: https://github.com/openstack/tempest/blob/master/tempest/scenario/test_volume_migrate_attached.py | 22:18 |
| smcginnis | xinli: But that creates and instance and attaches a volume and writes something like a timestamp on it. | 22:18 |
| smcginnis | xinli: Then detaches the volume, create a snapshot, create a new larger volume from that snapshot, and attach that new volume to the instance. | 22:19 |
| *** cknight1 has joined #openstack-cinder | 22:19 | |
| smcginnis | And within the instance assert that the timestamp is still there and that the volume shows up as a larger size. | 22:19 |
| smcginnis | If we had a tempest test like that, that would highlight which drivers really are not supporting this right. | 22:20 |
| smcginnis | And they would be forced to fix it since their CI would fail on every patch once this tempest test is merged to tempest. | 22:20 |
| smcginnis | Unless of course they skip the test, but we can check on that. | 22:20 |
| *** alonmarx has quit IRC | 22:21 | |
| smcginnis | Just an idea for now if you are highly motivated. :) Otherwise we can wait and see if there are any better ideas. | 22:21 |
| *** cknight has quit IRC | 22:21 | |
| xinli | @smcginnis: it seems that the erlon's generic solution won't be used in this case? | 22:22 |
| smcginnis | xinli: Well, maybe, maybe not. I think there is value in having a tempest test like that either way. | 22:22 |
| *** laurent\ has joined #openstack-cinder | 22:23 | |
| openstackgerrit | christopher.uhler proposed openstack/os-brick master: os-brick connector for Veritas HyperScale https://review.openstack.org/442754 | 22:24 |
| *** lpetrut has quit IRC | 22:26 | |
| *** cknight1 has quit IRC | 22:28 | |
| *** Jack_Iv_ has quit IRC | 22:29 | |
| xinli | @smcginnis: if core team wants to split the big patch into diff. smaller one for each drivers, can I also pick some , since I am familiar with these code? | 22:30 |
| *** Jack_Iv has joined #openstack-cinder | 22:30 | |
| smcginnis | xinli: Yeah, if you would want to help with that effort, that would be great. You do have a good knowledge of that area now after all you've gone through. :) | 22:32 |
| *** Jack_Iv has quit IRC | 22:32 | |
| *** lucasxu has quit IRC | 22:36 | |
| *** lucasxu has joined #openstack-cinder | 22:40 | |
| *** catintheroof has quit IRC | 22:41 | |
| *** catintheroof has joined #openstack-cinder | 22:42 | |
| *** Jack_Iv has joined #openstack-cinder | 22:44 | |
| *** Jack_Iv has quit IRC | 22:45 | |
| *** catintheroof has quit IRC | 22:47 | |
| openstackgerrit | Xinli Guan proposed openstack/cinder master: Fix 'driver does not honor clone/snapshot' bugs https://review.openstack.org/336092 | 22:48 |
| *** dustins has quit IRC | 22:52 | |
| xinli | erlon: | 22:53 |
| *** sdague has quit IRC | 22:57 | |
| *** ducttape_ has quit IRC | 23:04 | |
| *** erlon has quit IRC | 23:05 | |
| *** ducttape_ has joined #openstack-cinder | 23:07 | |
| openstackgerrit | Eric Harney proposed openstack/cinder master: Tests: Use six.moves.urllib urlencode https://review.openstack.org/442769 | 23:07 |
| openstackgerrit | Eric Harney proposed openstack/cinder master: Bump pylint to 1.4.3 https://review.openstack.org/442857 | 23:11 |
| *** lucasxu has quit IRC | 23:12 | |
| *** mlakat has quit IRC | 23:15 | |
| *** edmondsw has quit IRC | 23:16 | |
| *** eharney has quit IRC | 23:16 | |
| *** ociuhandu has quit IRC | 23:17 | |
| *** pck_ has quit IRC | 23:18 | |
| *** grumpycatt has quit IRC | 23:18 | |
| *** leseb_ has quit IRC | 23:18 | |
| *** grumpycatt has joined #openstack-cinder | 23:18 | |
| *** pck has joined #openstack-cinder | 23:19 | |
| *** leseb_ has joined #openstack-cinder | 23:20 | |
| *** lamt has quit IRC | 23:23 | |
| *** grumpycatt has quit IRC | 23:25 | |
| *** grumpycatt has joined #openstack-cinder | 23:25 | |
| *** xyang1 has quit IRC | 23:27 | |
| *** s-shiono has joined #openstack-cinder | 23:31 | |
| *** catintheroof has joined #openstack-cinder | 23:32 | |
| *** xinli has quit IRC | 23:33 | |
| *** mriedem has quit IRC | 23:36 | |
| *** ociuhandu has joined #openstack-cinder | 23:40 | |
| *** alonmarx has joined #openstack-cinder | 23:43 | |
| *** edmondsw has joined #openstack-cinder | 23:52 | |
| *** ducttape_ has quit IRC | 23:55 | |
| *** ducttape_ has joined #openstack-cinder | 23:56 | |
| *** edmondsw has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!