Safely reduce size of PvE installation installed to RaidZ-1 ?

[potetpro]

Hello.

We currently are having problem with some hybrid-disks, as seen here:
https://forum.proxmox.com/threads/c...w-ops-mon-proxmox4-crashed.66804/#post-300118


And now, about once a month:

ZFS has finished a scrub:

   eid: 36
 class: scrub_finish
  host: proxmox4
  time: 2021-06-13 00:25:06+0200
  pool: rpool
 state: ONLINE
status: One or more devices has experienced an unrecoverable error.  An
    attempt was made to correct the error.  Applications are unaffected.
action: Determine if the device needs to be replaced, and clear the errors
    using 'zpool clear' or replace the device with 'zpool replace'.
   see: http://zfsonlinux.org/msg/ZFS-8000-9P
  scan: scrub repaired 0B in 0 days 00:00:54 with 0 errors on Sun Jun 13 00:25:06 2021
config:

    NAME                                       STATE     READ WRITE CKSUM
    rpool                                      ONLINE       0     0     0
      mirror-0                                 ONLINE       0     0     0
        ata-ST1000LX015-1U7172_WKP6X74F-part3  ONLINE       0     0     3
        ata-ST1000LX015-1U7172_WKP6X67R-part3  ONLINE       0     0     0

errors: No known data errors

The disks are 1TB, and we are thinking about shutting the server down, cloning each disk, to SSDs and then power up again.

But we are thinking of using 128GB SSDs as boot volumes.
What is the best way to reduce the PVE volume before cloning the disks?
Is this a safe way?

Any insight would be greatly appreciated

 

[potetpro]

If there is no way to reduce the Raiz-1 volume, would it work to just, power down the server, use dd to clone the drives to larger disks, and then power up the server with the 2 new disks?

 

[ph0x]

Dd'ing to larger disks is always possible, the other way round, not reliably.
If I were to undergo such a procedure, I would probably reinstall on the new disks and transfer all appropriate data.

 

[potetpro]

The server also hosts part of two Ceph volumes (4 OSDs), I have not tested to restore these in testing before we deployed.