Slow prune lock the bucket for long time

Hello everybody,

i have a 3 node cluster with 12 SAS disks each node, all nodes have 2 X 10 gigs.

I’m backing up with restic, currently about 150Tb of data.

The backup runs smoothly, in fact, I often have to slow down from the source.

I have a problem with prunes … it takes a lifetime and they often lock my buckets, so I can’t make backups anymore and in short, everything crashes.

The servers are not overloaded, can someone recommend me something?

Which restic version do you use? How long does prune take? Which backend?

Using restic to backup hundred+ TB to a single repository is uncharted territory. I’d recommend to split the data onto multiple repositories if possible.

hi 0.12.1-3
backend is cluster, 3 node with 12 SAS disk every node.

Yes we already split in many bucket, we have 1 bucket per user.

But prune is very slow

Can you give some examples of your prune outputs which are slow? Maybe there is some other issue which can be analyzed when reading the outputs…

Which version is that? I recognize the 0.12.1. But the -3 is nothing official. Where does that come from? What does restic version print.

What does “slow” mean in minutes/hours? How large are the repositories? Please also provide the output of a slow prune run.