Ubuntu package versions

I have started upgrading my sensors to ubuntu 20.04 and found that puppet would not install suricata insisting that the package did not exist.

What I foudn was that on 18.04 package is version 6.0.1-0ubuntu2 but
on 20.04 it is 6.0.1-0ubuntu0.

I assume the \dubuntu\d refers to the build number and so can logically be different form OS version to OS version but I was wondering if there was a simple way of coordinating the package names across the OS releases.

Think they need to differ as PPA would not allow them to be uploaded to the PPA to be build if the same pkg name/version already exists.I would double check.
You mean more like give it an OS name (focal/bionic) instead of ubuntu ?

Hey Peter,

I thought there would be a reason. It isn’t a big deal and I have a somewhat messy work around but I thought I would ask. Changing ‘ubuntu’ to something related to the release won’t help me.

If you can double check that would be appreciated!

Most of my problem come back to the fact that I manage my sensors via puppet and decisions like exact version names have to be resolved on the puppet server without direct reference to the host. I.e. I can’t look and see which versions are in a particular repository. My work around is in my configuration json that gets fed into puppet I have replace build string with a hash, keyed by os.major_version.

Odd that I have not come across this before as I have 16.04 and 18.04 machines for at least a year. I am now upgrading the 16 to 20.

I could simply tell puppet to use latest but then I would loose control over the version entirely.

Running some checks - it seems I think i should be able to fix it. Will update.

I think this might be related to version numbers mapped to OS. Like higher version number for a pkg in relation to the OS. (ex 0buntu1 for 20.04 and 0ubuntu0 for 18.04)
Any way you could confirm that from previous installs ?

Forgot to mention - i can spin up some test pkgs if you want to try it out.

Thanks for the offer Peter but I think I will pass. I have a workaround in place and I am absolutely flat out!

I think it is just the case that this is the first time things have got out of step when I had multiple OS versions on my sensors. Can’t think of any way of verifying that historically.

Thanks for bringing it up and the feedback - always appreciated!

You are welcome Peter and thank you (and OISF) for your wonderful support!

It is highly likely that I will now be running multiple OSs as a matter of course.