| opendevreview | Merged openstack/glance master: Remove usage of eventlet for image cache SQLite timeouts https://review.opendev.org/c/openstack/glance/+/984314 | 00:56 |
|---|---|---|
| opendevreview | Sahil Kumbhar proposed openstack/glance_store master: S3 store: store location without credentials in the URI https://review.opendev.org/c/openstack/glance_store/+/988383 | 05:50 |
| opendevreview | Sahil Kumbhar proposed openstack/glance master: S3 store: store location without credentials in the URI https://review.opendev.org/c/openstack/glance/+/988385 | 05:59 |
| opendevreview | Sahil Kumbhar proposed openstack/glance master: S3 store: store location without credentials in the URI https://review.opendev.org/c/openstack/glance/+/988385 | 06:25 |
| opendevreview | Sahil Kumbhar proposed openstack/glance_store master: S3 store: store location without credentials in the URI https://review.opendev.org/c/openstack/glance_store/+/988383 | 06:27 |
| opendevreview | Abhishek Kekane proposed openstack/glance-specs master: [Spec] Parallel image import https://review.opendev.org/c/openstack/glance-specs/+/967046 | 06:42 |
| opendevreview | Abhishek Kekane proposed openstack/python-glanceclient master: Add support for GET /v2/cache/nodes/{image_id} https://review.opendev.org/c/openstack/python-glanceclient/+/988386 | 07:04 |
| opendevreview | Francesco Pantano proposed openstack/glance_store master: DNM: enable GLANCE_EC_ENABLE in ceph-tempest job https://review.opendev.org/c/openstack/glance_store/+/988421 | 09:58 |
| opendevreview | Takashi Kajinami proposed openstack/glance master: Use systemctl command to manage service https://review.opendev.org/c/openstack/glance/+/988460 | 13:46 |
| dansmith | mhen: I wonder if you could comment on the LUKSInspector patch indicating that you've tested it successfully and that it's needed for the now-approved glance image encryption effort? | 14:05 |
| dansmith | I think that might help spur some movement on review | 14:05 |
| dansmith | for reference: https://review.opendev.org/c/openstack/oslo.utils/+/978097 | 14:05 |
| dansmith | abhishekk: I think storing locations as pending is a good option if we can.. even just providing a manual cleanup command in glance-manage seems fine as long as we can identify them | 14:46 |
| dansmith | abhishekk: does the API show those locations or hide any that are not active? | 14:46 |
| abhishekk | get call shows the locations if show_direct_image_url or one other config option is set to True | 14:47 |
| abhishekk | but we can modify it to not show if the location is not active | 14:47 |
| dansmith | right, but does it show _all_ of them regardless of status? | 14:47 |
| abhishekk | yes | 14:47 |
| abhishekk | I need to check but I think it does show | 14:47 |
| dansmith | okay so filter out "pending" or something by default, maybe (later?) add an option to show pending as well? | 14:47 |
| dansmith | especially with parallel upload it might be nice to see which ones are done, which ones are queued | 14:48 |
| dansmith | waiting->uploading->active would also be nice for someone watching | 14:48 |
| dansmith | uploading->failed for the "not all must complete" would also be nice | 14:49 |
| abhishekk | ack, i will see how to fit this in the proposal | 14:49 |
| dansmith | I don't think you need to increase the surface area of this spec to include all that, just keep it in mind and pick a state that would work for a scheme like that, and filter it out of show for now | 14:50 |
| abhishekk | yes, I will just explain it in short | 14:50 |
| dansmith | cool "for future improvement" :) | 14:50 |
| abhishekk | thanks for the direction :) | 14:51 |
| dansmith | np | 14:51 |
| abhishekk | it loads active locations from the db only; https://github.com/openstack/glance/blob/master/glance/db/__init__.py#L103 | 15:07 |
| dansmith | aha cool | 15:08 |
| opendevreview | Francesco Pantano proposed openstack/glance_store master: DNM: enable GLANCE_EC_ENABLE in ceph-tempest job https://review.opendev.org/c/openstack/glance_store/+/988421 | 15:20 |
| croelandt | abhishekk: dansmith: so I think parallel image upload is the last thing we had on our plate for M1 | 15:32 |
| dansmith | cool, I think we have a plan for it | 15:32 |
| croelandt | yeah I might not be around to review the latest spec though | 15:33 |
| croelandt | as I'll be back on June, 1st :) | 15:33 |
| dansmith | jeeeeez | 15:33 |
| croelandt | French bank holidays, recharge, PTO about to expire... | 15:34 |
| dansmith | kinda one of the problems with glance's "only the PTL can approve specs" policy I guess | 15:34 |
| croelandt | We can approve it in June and not tell anyone :) | 15:34 |
| dansmith | I dunno what the actual gerrit rules look like but is it easy to add someone else for this one thing? | 15:34 |
| dansmith | that was my other suggestion.. | 15:35 |
| dansmith | comment on it that it's ready, implementation starting, delaying actual merge until the PTL returns from epicpto | 15:35 |
| croelandt | yeah if abhishekk improves the spec so we have an actual pool | 15:35 |
| croelandt | I'm fine with having the patch up and merging the spec in June | 15:35 |
| dansmith | ack | 15:35 |
| croelandt | abhishekk: I don't think we want to tag anything with M1? | 15:38 |
| croelandt | I don't think we had patches that are so amazing we want a new release ASAP | 15:38 |
| croelandt | Well, we've got "list cached nodes" | 15:38 |
| croelandt | And regarding the next 3 Thursday meetings, I'll let you all self-organize as if this were an anarchist Commune. Feel free to cancel them or keep them :) | 15:39 |
| abhishekk | croelandt: I think we are good for m1, I still need some time to update the parallel image import spec | 18:18 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!