Bug 1649679
| Summary: | Glance doesn't clean up the staging area with default config | ||
|---|---|---|---|
| Product: | Red Hat OpenStack | Reporter: | Thomas Hervé <therve> |
| Component: | openstack-glance | Assignee: | Abhishek Kekane <akekane> |
| Status: | CLOSED CURRENTRELEASE | QA Contact: | Mike Abrams <mabrams> |
| Severity: | medium | Docs Contact: | Kim Nylander <knylande> |
| Priority: | medium | ||
| Version: | 14.0 (Rocky) | CC: | akekane, eglynn, gfidente, joflynn, tenobreg |
| Target Milestone: | --- | Keywords: | Triaged, ZStream |
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Unspecified | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Release Note | |
| Doc Text: |
When you use the web-download feature, the staging area - defined in the configuration using the `node_staging_uri` option - is not cleaned up properly. Ensure that `file` is part of the `stores` configuration option in the `glance_store` section of the glance-api.conf file.
|
Story Points: | --- |
| Clone Of: | Environment: | ||
| Last Closed: | 2020-05-05 09:38:16 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
It looks like it was present since the beginning, too: https://bugzilla.redhat.com/show_bug.cgi?id=1418436#c8 So, I talked to Thomas. If he adds "file" to the list of stores in glance-api.conf, everything works as expected. I do not think that this is especially user-friendly, though. Shouldn't we automagically load the "file" store? This could be done in the _DeleteFromFS class: https://github.com/openstack/glance/blob/master/glance/async_/flows/api_image_import.py#L78 We already use this workaround here: https://github.com/openstack/glance/blob/master/glance/async_/flows/api_image_import.py#L149 @Abhishek: do you think upstream would be ok with this? Hi Cyril, Thomas, I have proposed a upstream fix for this issue. Please refer, https://review.openstack.org/618468 |
I tested the web-download import with glance on containers, and saw the following error in the glance logs: Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/taskflow/engines/action_engine/executor.py", line 53, in _execute_task result = task.execute(**arguments) File "/usr/lib/python2.7/site-packages/glance/async_/flows/api_image_import.py", line 92, in execute store_api.delete_from_backend(file_path) File "/usr/lib/python2.7/site-packages/glance_store/backend.py", line 409, in delete_from_backend loc = location.get_location_from_uri(uri, conf=CONF) File "/usr/lib/python2.7/site-packages/glance_store/location.py", line 75, in get_location_from_uri raise exceptions.UnknownScheme(scheme=pieces.scheme) UnknownScheme: Unknown scheme 'file' found in URI The import worked though, and the image seems active. I have the following defined the config: node_staging_uri=file:///var/lib/glance/staging It seems this area doesn't get cleaned up, and will get clutter over time. I used glance in the undercloud with containers, but I don't think it would work in the overcloud either.