Learn more about these different git repos.
Other Git URLs
Describe the issue Enabling the Cisco repo by default in Fedora 32 Workstation results in a dnf error. We think the best approach is to use MirrorManager to handle the repo metadata. This request enables that as described in BZ 1768206c24
When do you need this? (2020/04/16)
If we cannot complete your request, what is the impact? A release blocking bug is not resolved.
opening on behalf of @adrian
I just saw that it would also be possible to trigger the sync from mm-backend01 as mm-backend01 already has access to the proxies via the mirrormanager user.
The main question for me right now is from where is the sync triggered. It would also be good to know how often MirrorManager has to scan for changes or if we can also trigger the scan of this repository (umdl) somehow.
So, @mohanboddu and I discussed this today and thought:
Short term (since we are going into freeze and since the h264 repo is currently manually created, but will be made by ODCS hopefully later) we just manually sync the repodata to mm-backend01 whenever we update the repo. Where would be good for this? /srv/web/whatever ? We could at the same time trigger mm to re-read it? So, it would need an initial setup to see it, but then after that we only tell it to update when we update the repodata.
Longer term after the release and the datacenter move and ODCS being used, we setup something more automated.
Sound ok? What dir makes sense to put the repodata? I guess we can sync the current repodata from the current live repo. ;)
Thoughts?
Sorry, I missed your question:
The way it works now is:
For now I have also synced the content to /srv/codecs.fedoraproject.org/ on mm-backend01 for you.
It seems to work:
$ curl -s "https://mirrors.fedoraproject.org/metalink?repo=fedora-cisco-openh264-32&arch=x86_64" <?xml version="1.0" encoding="utf-8"?> <metalink version="3.0" xmlns="http://www.metalinker.org/" type="dynamic" pubdate="Thu, 09 Apr 2020 10:22:49 GMT" generator="mirrormanager" xmlns:mm0="http://fedorahosted.org/mirrormanager"> <files> <file name="repomd.xml"> <mm0:timestamp>1584475845</mm0:timestamp> <size>3074</size> <verification> <hash type="md5">e2f94cd477d08d8c05a435fcdb2d723d</hash> <hash type="sha1">26cb3a40f8f828769d8fed09af7d0631c74573e7</hash> <hash type="sha256">5c95a431fa860c69867cbdff65d2e3506c9ca6cb4670223d40cdb3e657487c2a</hash> <hash type="sha512">de714821454a94c6e6be3ce765ffa2764c8b458c3f625c9db7e904ccfbd7cca93b02995873473d893e8a30a3868e57cac83f1db1ea4179f985ee484640f94321</hash> </verification> <resources maxconnections="1"> <url protocol="https" type="https" location="US" preference="100">https://codecs.fedoraproject.org/openh264/32/x86_64/repodata/repomd.xml</url> </resources> </file> </files> </metalink> $ curl -s https://codecs.fedoraproject.org/openh264/32/x86_64/repodata/repomd.xml | sha256sum 5c95a431fa860c69867cbdff65d2e3506c9ca6cb4670223d40cdb3e657487c2a -
As always it required small code changes for the repository detection, which I will soon bring upstream.
From my point of view this should be usable/testable now.
I removed the baseurl line from my fedora-cisco-openh264.repo file and added: metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-cisco-openh264-$releasever&arch=$basearch
metalink=https://mirrors.fedoraproject.org/metalink?repo=fedora-cisco-openh264-$releasever&arch=$basearch
and a dnf reinstall openh264 worked (F32, x86_64).
dnf reinstall openh264
We still need to figure out how this can be automated. Right now even UMDL has to be run manually.
fedora-repos pull requests to switch to metalink:
https://src.fedoraproject.org/rpms/fedora-repos/pull-request/57 (rawhide) https://src.fedoraproject.org/rpms/fedora-repos/pull-request/58 (f32)
Upstream MM changes: https://github.com/fedora-infra/mirrormanager2/pull/282
The work is done and this can be closed now.
Metadata Update from @mohanboddu: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Login to comment on this ticket.