| opendevreview | Tim Burke proposed openstack/swift master: sq: cross-compat test single-part obj partNumber and Range https://review.opendev.org/c/openstack/swift/+/906906 | 00:01 |
|---|---|---|
| opendevreview | Tim Burke proposed openstack/swift master: slo: Support range requests for part-number queries https://review.opendev.org/c/openstack/swift/+/906391 | 00:01 |
| opendevreview | Merged openstack/swift master: Prevent installation of known-broken eventlet https://review.opendev.org/c/openstack/swift/+/906435 | 01:55 |
| zigo | Hi there! | 09:10 |
| zigo | I'd like to propose a patch to add (a new version of) the script to check for device full and modify rsyncd.conf at https://docs.openstack.org/swift/latest/admin_guide.html#preventing-disk-full-scenarios | 09:10 |
| zigo | It's tkajinam that's suggesting it'd be nicer to have it in swift directly rather than in puppet-swift. What's the team's view on this? Will such a patch be accepted? | 09:11 |
| tkajinam | we already have swift-drive-audit in swift then we can probably add something like swift-drive-space-audit (or even add this feature to swift-drive-audit ?) | 09:19 |
| opendevreview | Merged openstack/swift master: proxy: only use listing shards cache for 'auto' listings https://review.opendev.org/c/openstack/swift/+/901335 | 11:59 |
| timburke | zigo, tkajinam: seems reasonable -- there might be a few pitfalls to watch out for when writing a more serious script, like unmounted devices, but i like the idea of more operator tooling living upstream | 18:32 |
| timburke | 🤔 so on https://bugs.launchpad.net/python-swiftclient/+bug/1541018 -- how *do* we want to handle relative paths during uploads? if i run `swift upload test some/path/../../no/i/meant/here`, should it upload to $OS_STORAGE_URL/test/some/path/%2E%2E/%2E%2E/no/i/meant/here or to $OS_STORAGE_URL/test/no/i/meant/here ? | 19:19 |
| patch-bot | Bug #1541018 - Uploading a file from parent dir with relative path results in two containers created and no file upload (In Progress) | 19:19 |
| timburke | does the answer change if i run `swift upload test no/i/meant/here --object-name some/path/../../no/i/meant/here` instead? | 19:19 |
| timburke | i *think* https://bugs.launchpad.net/python-swiftclient/+bug/2051978 might be separate but related -- for the one bug, we're taking actual filesystem paths as input and trying to map that to an object name, while the other is all about trying to work within the object namespace (though there's a similar translation question for downloads) | 19:23 |
| patch-bot | Bug #2051978 - Cannot download, delete, or stat paths with . or .. as components (New) | 19:23 |
| zigo | timburke: We have already some kind of script that's using configobj to parse /etc/rsyncd.conf, we just need to finish and release it. | 20:59 |
| zigo | Hopefully, we can send the PR next week. | 21:00 |
| timburke | nice | 21:00 |
| opendevreview | Anish Kachinthaya proposed openstack/swift master: added a new line to readme https://review.opendev.org/c/openstack/swift/+/907136 | 23:47 |
| timburke | new intern, checking out the git-review mechanics ;-) | 23:50 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!