[SOLVED] radosgw install (attempt) odd error messages

[SpeakEZ.ai]

I'm using this page as a guide for adding radosgw to Ceph in order to run S3 APIs in a dedicated pool. This is the error I get when running "apt install radosgw"

> apt install radosgw
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
Some packages could not be installed. This may mean that you have
requested an impossible situation or if you are using the unstable
distribution that some required packages have not yet been created
or been moved out of Incoming.
The following information may help to resolve the situation:

The following packages have unmet dependencies:
 radosgw : Depends: ceph-common (= 17.2.7-pve3) but 18.2.2-pve1 is to be installed
           Depends: librgw2 (= 17.2.7-pve3) but 18.2.2-pve1 is to be installed
E: Unable to correct problems, you have held broken packages.

The *interesting* bit is when I check the version of ceph-common and librgw2 they both show 18.2.2.

> dpkg-query -f '${Version}' -W ceph-common
18.2.2-pve1
> dpkg-query -f '${Version}' -W librgw2
18.2.2-pve1

Perhaps I'm misinterpreting what the error message ("is to be installed"?) is telling me. Those versions *are* installed. Is it a matter of radosgw *looking* for the previous version? Can anyone help with pointers on getting through this issue? Thanks!

 

[VictorSTS]

The version of radosgw that the system is trying to install seems to be 17.2.7, but your currently installed versions of other Ceph packages seem to be 18.2.2.

Are your sure the repositories are correctly configured in that system?
Did you run apt update?
What's the output of apt-cache policy radosgw?

 

[SpeakEZ.ai]

Thanks for echoing back - I'm not sure of anything since I started seeing those errors.

Here's the output

> apt-cache policy radosgw
radosgw:
  Installed: (none)
  Candidate: 17.2.7-pve3
  Version table:
     17.2.7-pve3 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     17.2.7-pve2 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     17.2.7-pve1 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     17.2.6-pve1+3 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     16.2.11+ds-2 500
        500 http://ftp.us.debian.org/debian bookworm/main amd64 Packages

FWIW this is my output of apt update

> apt update
Get:1 http://security.debian.org bookworm-security InRelease [48.0 kB]
Hit:2 http://ftp.us.debian.org/debian bookworm InRelease
Get:3 http://ftp.us.debian.org/debian bookworm-updates InRelease [55.4 kB]
Get:4 https://packages.microsoft.com/debian/12/prod bookworm InRelease [3,618 B]
Hit:5 https://enterprise.proxmox.com/debian/ceph-quincy bookworm InRelease
Hit:6 https://enterprise.proxmox.com/debian/pve bookworm InRelease
Fetched 107 kB in 6s (17.8 kB/s)
Reading package lists... Done
Building dependency tree... Done
Reading state information... Done
All packages are up to date.

 

[VictorSTS]

> apt update
[...]
Hit:5 https://enterprise.proxmox.com/debian/ceph-quincy bookworm InRelease

You have Ceph Quincy (17.x) repo configured instead of Reef (18.x).

Curious about how you may have some Ceph Reef packages installed, what's the output of apt-cache policy ceph-common?

 

[SpeakEZ.ai]

I appreciate you helping me figure out this mystery.

> apt-cache policy ceph-common
ceph-common:
  Installed: 18.2.2-pve1
  Candidate: 18.2.2-pve1
  Version table:
 *** 18.2.2-pve1 100
        100 /var/lib/dpkg/status
     17.2.7-pve3 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     17.2.7-pve2 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     17.2.7-pve1 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     17.2.6-pve1+3 500
        500 https://enterprise.proxmox.com/debian/ceph-quincy bookworm/enterprise amd64 Packages
     16.2.11+ds-2 500
        500 http://ftp.us.debian.org/debian bookworm/main amd64 Packages

I'm honestly not sure how Quincy worked its way into the mix - or how I could have prevented it. I haven't put my work/data load on this cluster yet as block storage is one of the underpinnings I need in the lab. So while I'm not eager to wipe the Ceph setup and "start from zero" I can do it without risk to any data.

 

[VictorSTS]

Seems like at some point you had Reef repo configured but now you have Quincy: the ceph-common package that you have currently installed is not available in any of your configured repos.

Simply remove the Quincy repo and add the Reef one.

 

[SpeakEZ.ai]

I'm grateful for the advice. I'm honestly not certain how it got into this state. I ran the Ceph install on all four machines that are supporting it, and *thought* I had selected the same options - but one is correct and three are not. So I have some clean-up to do.

My sources.list is pretty skinny...

deb http://ftp.us.debian.org/debian bookworm main contrib

deb http://ftp.us.debian.org/debian bookworm-updates main contrib

# security updates
deb http://security.debian.org bookworm-security main contrib

I'm scouting around for the list of repos I'm seeing in the apt-cache command but coming up empty. Can get just a *bit* more help?

 

[VictorSTS]

Use the webUI, it's quite easy:

- go to the host, updates -> repositories. Here check if you have any Ceph repository (will be configured in /etc/apt/sources.list.d/ceph.list).
- If there's any, make sure it's the "Reef" one.
- If there's a Quincy one, click on it and press "Disable" button. Then add the "Reef" one. Make sure you dont have both!
- If there's none, simply add the "Reef" one.
- Then go to Updates, and click the Refresh button.
- Check with apt-cache policy radosgw that the Reef version is offered in the repo.
- Install the package.

Use no-subscription repo if you dont have one (but I think you do have subscription given previous information).

 

[SpeakEZ.ai]

OK - great - I was just mulling the idea of re-imaging the machines while eating lunch (US east coast time). I'm at the stage where "going back to move forward" is not that worst option. I'll check it out. Thanks again for helping a newb crib through this!

 

[SpeakEZ.ai]

As an update I *think* the issue is that installing from the command line uses Quincy while using the we UI uses Reef (by default).