Virtualmin APT update problems

No, the key was generated in 2022. Jammy was the current version then, and we have something like 75,000 active installs on Ubuntu versions from 20.04 through 24.04. This key definitely works with Jammy (and Focal and Noble).

Legit. That also makes sense, since the keyring ACCEPTED the key. Both apt-key and gpg --list-keys --keyring ... show the key and don’t complain about it at all.

But apt itself simply balks at it.

ETA: For funzies, I also went old-school and used apt-get instead of just apt. Same result.

It’s a known, tricky, intermittent issue on Ubuntu that’s not directly tied to Virtualmin but affects many users.

Please refer to this thread:

If you will read uptopic, you will see that I have already gone through the steps outlined in that thread, which are also the steps recommended to me separately, here, and that exactly none of them made any difference.

Just for fun though, I went through the suggestion from that thread:

ALTERNATE METHOD:

sudo gpg --keyserver pgpkeys.mit.edu --recv-key  <PUBKEY>
sudo gpg -a --export <PUBKEY> | sudo apt-key add -
sudo apt-get update

With the following result on the first command:

root@bee:/home/mikey# sudo gpg --keyserver pgpkeys.mit.edu --recv-key 3E570892B9A0B8B7
gpg: keyserver receive failed: No data

We haven’t published our key to pcpkeys.mit.edu. That can’t possibly do anything.

Since we are grasping at straws, force a reinstall of apt?

I wouldn’t have thought so, either, but @Ilia seemed to think I should try it, so I tried it!

Removed… Sorry I anticipated to help out.

As you can see above, I am root when I am executing these commands.

If apt were failing for anything else, I’d think that might help, but all the other repos apt is pulling from work fine. I can update Ubuntu’s own packages without an issue, and in fact just did so a short while ago.

Removed… Sorry I anticipated to help out.

I have been administering Unix and Unix-like systems for 35 years. I have never once, in all that time since sudo came into existence at all, heard of anything at all where sudo -s behaved different from logging in as root.

Nevertheless, since we are, as @ID10T points out, down to grasping at straws, I tried logging in as root, and running virtualmin setup-repos again.

It made no difference whatsoever.

Since you know it all with that attitude… Like I said… Best of luck!

Look, folks.

There are two broad possibilities.

Either apt has a bug, or Virtualmin’s repo has a problem.

Since nothing else is broken, only virtualmin’s repo; and it’s consistently broken no matter what remedy I apply, including getting rid of every possible reference to it and reinstalling with Virtualmin’s own installer script…that makes it seem to me that it’s Virtualmin’s problem.

I note that someone else in this community had this same problem all the way back in April, and the only answer was shifting the blame to the installer for a ā€œcreativeā€ installation and insisting he would not have had this problem if he’d just followed the instructions.

Well, I’ve followed the instructions, and it’s still broken.

I’m not trying to argue with you about it. I just don’t know what’s wrong. I can’t reproduce it, I’ve tried, and I’m still trying to figure out what could be going wrong.

The only theory I have is that there’s something about apt configuration on systems that have been upgraded from a quite old version of Ubuntu (ten years would be trusty, maybe?). So far I haven’t found confirmation of that, but that’s the best theory I have, since both people who have reported the problem were on systems that had been upgraded, rather than a fresh install.

I’m not criticizing anything you’ve done, nor am I saying this is your fault.

@Joe I appreciate that. And L-rd knows I’ve been in your shoes enough times, since I’m also a software engineer by day, and sometimes, systems just get…weird.

The only theory I have is that there’s something about apt configuration on systems that have been upgraded from a quite old version of Ubuntu (ten years would be trusty, maybe?).

That theory holds water. I actually just delved into the droplet’s ā€œhistoryā€ page, which strangely does not include a specific date, but does confirm it was created 10 years ago, which was honestly just a ballpark when I first mentioned it. The services the system is running date back further than that – I ran a small webhosting business that went from hosted-at-home (frac-T1 to the house of all things) to colo to Digital Ocean. The system in question is the remaining rump of that business, just a handful of vhosts for friends and a couple of things of my own.

I’m not sure what to look for to fix it–there’s been nothing glaring when I looked at /etc/apt’s config files–but it’s close to the last sensible thing we have left!

My point was that maybe some configuration got left, so forcing a complete reinstall might help. I can’t do a large upgrade without pauses making sure I know the consequences of something or having it offer me an option to keep a configuration or go with the packagers new version.

Time to start a fresh on a new box - perhaps?

It can’t, as long as we believe the Virtualmin key is the real issue. Again, I know that apt thinks it is, but it could be a bug. However, I have no solid evidence to support this.

Was this system upgraded from 20.04 or 18.04? And were the upgrades done in consecutive years or one after the other a few days ago?

Ten years ago would be starting with 12.04 or 14.04.