Andrew Sauber
Andrew Sauber
The word "enforce" to me usually implies a security option or functionality restriction. I think a better name would be `--force-device-detection`.
Unassigning myself as I've shifted to projects and don't want to block this.
In the draft implementation, there is a 30 second delay added to any CID deletion-via-operator here https://github.com/cilium/cilium/pull/33204/files#diff-a5b01710162b5dcb0c820096bd611ded7a00e14a8a60ebc382466ce30d9436caR249-R253 To me, this appears to introduce an [at least] 30 second delay between...
Apologies for spam. The relevant diff is not linkable on GitHub due to the diff being collapsed by default. > In the draft implementation, there is a 30 second delay...
I also lost access to one of my machines. I use `authorizedkeyscommand` in the `/etc/ssh/sshd_config` file. This change https://github.com/systemd/systemd/pull/30061 overwrote that setting, and I could no longer access my machine...
> That sounds like a packaging bug. The package should not install the symlink without the file. The issue is not whether or not the symlink is installed correctly. The...
Another issue here is users who are not using AuthorizedKeysCommand or sssd at all. Users who are familiar with sshd expect that by default the only way that a keypair...
Hi all, my apologies, I did misread the original diff. I have dropped an approval with a suggestion for changes to the description to match the surrounding descriptions. Thank you...
We have had some feedback that `cilium status` takes too long to execute for large clusters. This seems potentially quite useful, but we would have to keep performance in mind.
Imposing a cutoff time for the response from all agents is a great idea.