It seems that our CI runs are now failing because a CentOS 6 repo disappeared:
Status: Downloaded newer image for centos:6
dc404c3f71e249ccce98ddd5dbf9c7a7ad976beef9e13b5dd3d7cd61eec1e441
Loaded plugins: fastestmirror, ovl
Setting up Install Process
Error: Cannot retrieve repository metadata (repomd.xml) for repository: base. Please verify its path and try again
YumRepo Error: All mirror URLs are not using ftp, http[s] or file.
Eg. Invalid release/repo/arch combination/
removing mirrorlist with no valid mirrors: /var/cache/yum/x86_64/6/base/mirrorlist.txt
child_process.js:660
throw err;
^
Error: Command failed: docker exec centos yum install -y centos-release-scl cmake xz epel-release
("Cannot retrieve repository metadata" seems to be the key bit.)
It seems that CentOS 6's EOL was last month (Nov 2020): https://phoenixnap.com/kb/centos-6-eol. I imagine that someone just pushed the button to delete the repo.
Should we move to CentOS 7 for CI builds?
ah, interesting! We're using that old version to ensure compatibility with old versions of Linux more generally, so I guess the question is whether we want to change how far back we go with that. @Alex Crichton, wdyt is the best way forward here?
oh dear
We should probably bump to centos 7 and only worry about fixing centos 6 if someone files an issue
I'm about to create a PR to pull us to CentOS 7
seems to be just minor tweaks
I think we can get 6 working but it would be a pain
let's see if this becomes green: https://github.com/bytecodealliance/wasmtime/pull/2464
agreed: moving to CentOS 7 makes sense
Last updated: Dec 23 2024 at 14:03 UTC