Following the expiration of a large number of backups, the AVE disks are very full:

Filesystem Size Used Avail Use% Mounted on
/dev/sdb1 1000G 974G 26G 98% /data01
/dev/sdc1 1000G 974G 27G 98% /data02
/dev/sdd1 1000G 971G 29G 98% /data03

The problem started on the 31st of May:

 

But now we are in a chicken and egg situation. It won't checkpoint and it won't garbagecollect. HFS check won't run because it needs a new checkpoint. Not that it'd save space anyway...

Checkpoint failed with result MSG_ERR_DISKFULL : cp.20210603145550 started Thu Jun 3 15:56:20 2021 ended Thu Jun 3 15:56:20 2021, completed 0 of 6509 stripes
Last GC: finished Thu Jun 3 16:01:44 2021 after 00m 30s >> recovered 0.00 KB (MSG_ERR_DISKFULL)
Last hfscheck: finished Mon May 31 09:41:56 2021 after 01h 02m >> checked 5837 of 5837 stripes (OK)

So what can one do now? I have deleted all but the one remaining validated checkpoint:

Tag               Time                    Validate  Deletable
----------------- ----------------------- --------- ---------
cp.20210531065414 2021-05-31 07:54:14 BST Validated No

I wish there was a way to trash all checkpoints (including the current one) without affecting client configurations... I don't care about the backed up content anyway.