How to backup a Ceph Pool via PBS (Not VMs)

[MGSteve]

Hi All,

Background
We currently have a CRM which is written in PHP, running on the LAMP stack. We were using HyperV, but I managed to talk the boss into getting a new platform, 3 PBE nodes and a PBS backup server - All physical servers, not cloud etc..

As part of the CRM we store millions of files (Emails, documents etc..), which is currently handled by a single VM on the Hyper-V setup acting as a fileserver via NFS. it works OK - file until now when the disk is running out of space and the HyperV Server itself has no more disk space available - one of the reasons for the re-platform (the Hyper-V servers are old too).

The Question
You may have seen a similar thread I just posted about an issue with CephFS I ran into, upon googling around I realised that there was a PHP interface available to Ceph as a document store - I had been thinking of re-engineering the app to store these files in a document store for a while - they were originally stored in the mySQL database (that was done before my time), one of the first things I did was to strip them out into the Filesystem. Back then there were only about 3m files!

Anyway, that got me thinking - that's fine & all storing them in Ceph via the document store, but how do you back that pool up? Can you backup a Ceph Pool via PBS?

The plan would be to setup a new Ceph Pool and just use that for the document store, so there's a logical separation from the VM & other proxmox storage. - I assume that's how I'd do it anyway?

Any help would be appreciated

Thanks

 

[sb-jw]

So CEPH is basically a block storage. It now also offers CephFS, which is a file system. You can also create object storage using Swift and S3.

Now the question is, where exactly do you want to save the files?

But in general the answer to your question is that it is possible. This usually works if you have a VM or server that has mounted the CephFS, for example, on which you can then explicitly save this mount using the PBS client. The same applies to block or object storage. You won't be able to backup this directly from CEPH.
However, this is not a limitation of CEPH, it is basically the case that block storage cannot be backed up directly.

 

[MGSteve]

So CEPH is basically a block storage. It now also offers CephFS, which is a file system. You can also create object storage using Swift and S3.

Now the question is, where exactly do you want to save the files?

But in general the answer to your question is that it is possible. This usually works if you have a VM or server that has mounted the CephFS, for example, on which you can then explicitly save this mount using the PBS client. The same applies to block or object storage. You won't be able to backup this directly from CEPH.
However, this is not a limitation of CEPH, it is basically the case that block storage cannot be backed up directly.

Thanks for the reply

Yeah, In my other post I did come across an issue with CephFS - which may well be a settings change I need to make in order for it to work. But as it stands, I can't copy the existing files into the CephFS instance without it failing.

 

[MGSteve]

Or is the simplest solution simply to keep with the fileserver VM approach, now I can make use of the HA features via PBE & Ceph and simply manage the VM backups via PBS?

It just seems "wrong" in 2024 to have a VM fileserver simply to provide access to the files - especially in the case of Hyper-V, when a restore of the VM took 11 hours!! (Hopefully that won't be the case with PBS).

Especially when there are purpose designed document databases (couchbase, MongoDB etc..) around, but none of which are integrated into PBS.