How to identify which pg's are assigned to which drive in a 3-node CEPH cluster?

[FSNaval]

Good day all,

I have a 3-node proxmox cluster with CEPH, using for data storage hdd drives, some of which i had them from laying around.

I think that some of them (1 or 2) are starting to kick the bucket, because i receive the attached warnings. How can i find out which pg's are assigned to which hdd so that i can identify which drive needs to be replaced?

 

[Maximiliano]

Hello,

I don't think those messages necessarily imply that the disks are failing or about to. Nevertheless you can see where each PG is with the following command:

 ceph pg dump

Could you please attach the full output of

ceph status

I would suggest to look for IO errors in the system journal or SMARTD before concluding that the disks have a hardware problems. In any case, it is known that HDD disk struggle to keep up with Ceph's IO, which mostly consists of random-read and random-write operations. We recommend fast enterprise grade SSD disk for Ceph [1].

[1] https://pve.proxmox.com/pve-docs/pve-admin-guide.html#_system_requirements