Epel 8 rhel 8 public repo for use for a foreman server "product"

scenario: I inherited the admin of a foreman installation, with stuff all docuementation from its previous admin. Im learning on my feet, slowly, as i go along so if anythuing that follows seems daft, or Foreman 101, thats why.

Problem:

Content - Products - epel_x86-64
Three Repositores within
epel7
epel8_x86-64
epel_rhel8_modular

first two of those show failures oin syncing. Im not overly bothered about epel7 as centos7/rhel7 is ouit of support now anyway.

epel8_x86-64
has an upstream URL of Index of /fedora-epel/8/Everything/x86_64/
verify SSL is set to Yes
there are no GPG keys or SSL client Cert/Keys defined.
Download policty is “Immediate” abnd Mirroring Policy is Content Only (if they are even relavant here).

If I attempt to run a “Sync Now” it takes a while as “running” but doesnt seem to get past 39%
I can see in the Running Steps the final line ““poll_attempts”=>{“total”=>29, “failed”=>0}}” while it is “running”, where that total increments every few seconds (about 15 seconds or so)

Eventually the task enters “stopped” after 8 minutes or so, and then shows 100% complete. With an orange coloured bar.

The error shown is “Fatal parser error” which is less than helpful …:slight_smile:
Running steps now shows “no running steps”

Errors shows “poll_attempts”=>{“total”=>49, “failed”=>1}}

and in

Exception: field = Katello::Errors::Pulp3Error: Fatal parser error
Action: field = Actions::Pulp3::Repository::Sync
Actions::Pulp3::Repository::SaveVersion *
Actions::Pulp3::Repository::CreatePublication *
Actions::Pulp3::Repository::SavePublication *
Actions::Pulp3::Repository::SaveDistributionReferences *
Actions::Pulp3::Orchestration::Repository::Sync *
Actions::Katello::Repository::ErrataMail *
Actions::Katello::Repository::Sync *
Actions::Katello::Repository::Sync *

  • all these also have a field
    Output:
    {}

Ive no idea whether any of that is even relevant.

If I try and load the configured URL Index of /fedora-epel/8/Everything/x86_64/ in a browser is loads absolutely fine ands gives me a directopry listing of

Index of /fedora-epel/8/Everything/x86_64/


Packages
debug
drpms
repodata

all those links also load absolutely fine. QED its not a broken link scenario. There are no expoired GPG keys etc as none have been/were ever configured previously.

Expected outcome:
Needed outcome is to be able to sync this repo whether from the opriginal URL or wherever else (usual caveats)

Foreman and Proxy versions:
Version 3.2.0

Foreman and Proxy plugin versions:
??

Distribution and version:
??

Other relevant data:
N/A though happy to supply if pointed to what I need to provide.

Foreman 3.2.0 is an old version and out of support. The latest released version is 3.13.

It maybe easier to set up a new foreman server with latest version.

This document would help if you want to install the latest Foreman.
https://docs.theforeman.org/

That is on the cards in 2025 (ie upgrade) but in the meantime Im tasked by the PTB to get this working now.

And whether there’s an upgrade or not available doesnt exp[ain why the same version worked for epel8 as described until about July but not since of course which is somewhat bizarre given the availability of the upstream URL is good via a browser for example.

Well, having logged in as the admin user and not me, and rerunning the sync I do see more info avbailable at least. Though ultimately the same fate!

I get the same response if I try another alleged upstream URL
https://dl.fedoraproject.org/pub/epel/8/Everything/x86_64/

which I gleaned from googling for “foreman uptream url epel8”.

No content added.
Total steps: 11421/11421

Associating Content: 1097/1097
Downloading Artifacts: 16/16
Downloading Metadata Files: 6/6
Parsed Packages: 10302/10302

Upgrading is essential. Dealing with issues which possibly have been solved in later versions doesn’t really help except taking up much time to work around until you update.

It does very well explain this, if there is a bug in an older version or something incompatible in your version which doesn’t deal properly with rpms packaged later.

You really have to provide the exact foreman and katello version you are using as well as giving your distribution and the production.log output for that exception.

And please use the proper blockquote/preformatted tags in your posts. They are really hard to read if you just mix output between your own text using the same formatting…

OK. Please advie exactkly how to provide “the exact foreman and katello version I am using as well as giving your distribution and the production.log output for that exception.”

As explained this has been dumped in my lap and the that used to administer this pile of crap is gone and all the knowledge of this went in his head. I’m flying blind and frankly googling elicits nothing that is clear and obvious to me.

Sorry to so useless.

Meanwhile regarding upgrading etc is there a nice bullet pointed “how to” somewhere out there on the web? I wouldn’t even know where to start.

I’m phuqqed basically.

You could always look at other people posts on which information to give. Don’t expect people here to give you a beginners introduction course into foreman. There are docs at https://docs.theforeman.org/ . It tells you where to find the logs and troubleshooting information. If you are serious you have to dig into the docs anyway, so there is little point reciting them here… It tells you how to upgrade.

And basic knowledge for your operating system would be helpful, too, e.g. to find the versions of rpms installed or where to find the name and version of your operating system. If you have problems finding and providing this basic information I am not sure if it’s a good idea to continue anyway. You need good linux knowledge to operate a foreman server anyway. Without out, you won’t get far and it won’t be any fun…

So as pointed out before: it may be the best for you to make a new installation. That way you learn everything you need to know right from the start.

yeah. And I’ve today to do that.

Good 'ere innit.

To be somewhat productive here: Fatal parser error (at least in my experience) hints at Pulp (the component that does the actualy repo syncing and management in the background) being unable to parse upstreams metadata. This is probably also the reason why you can browse the repos just fine via browser and why newer versions are not affected.
I have found this thread on the forums that mentions similar problem to yours and on a similarly old version. There is a patch linked the you can try to apply to your pulp installation (remember to have some form of backup at hand if this go wrong).

As a general hint at troubleshooting repository problems: Pulp by default logs to both /var/log/messages and Journal. The easiest way to see logs from Pulp would be journalctl -u 'pulp*'. Depending on how many hosts your Foreman manages, this can get quite noisy, but if you have the problem I expect (which is the one I linked above) you should find the “Fatal parser error” in there, followed by a stack trace with something about parsing metadata near the end.

1 Like

Still, you should try to update your system to at least a somewhat recent version asap, although I understand that’s quite a daunting task for someone new to Foreman. Hopefully the info from my post gives you enough breathing room to pursue the upgrade.

To get to a current version of your stack you basically have two options, as already mentioned:

  • The “classic” update, taking your current setup to a recent version. This has the advantage of being well tested and documented. You always have to do one minor version at a time (so updates like 3.2 → 3.3 → 3.4 → etc until you hit a current version). You can find the upgrade steps for each version in the docs (as already mentioned). Though, the downside to this approach is that if something goes wrong, it’s your production system that stays down. Depending on the age of your installation and how well maintained any additional plugins you use are, this might be more or less likely. You will usually be able to find a solution to any upgrade problems searching the forums, but this of course costs additional time and nerves.
  • The second option (as already mentioned by lfu) would be to setup a new system with a current version and then move your hosts over to the new Foreman stack. Upside to this approach is that you can directly start with a clean system that you know and are not risking your production system. The clear downside is that this approach takes way longer and there is very little guidance available, because the steps you need to take are primarily determined by which features of Foreman you are using.
1 Like

Thanks Areyus for your positive help. I confess I was hoping someone may say "oh - the epel 8 repos got moved to but as ever in life its never that simple :slight_smile:
cheers

I suspect it will be a new system cos what I have inherited is on CentOS7 and needs an OS upgrade anyway, hence my earlier mention that an upgrade is in the pipeline.

the problem I have as you have understood is that given the last tw4t wrote the square root of fwck all down I’m flying blind as to how to install foreman, how to configure it as he did and what “extras” may need doing alongside it.

cheers