I thought this would be interresting for others too:
If you want to use a central PBS for more then one cluster, you run into the problem of double IDs.
Ok, you can use different datastores, but then you have to know the expected sizes.
We did an other aproach:
We build one big datastore and afterwards we add datastores via absolute backing path which are below the big datastore.
Datastore data-1 (the main ZFS datastore) is located in:
/mnt/datastore/data-1
So I added a datastore user-1 with the path:
/mnt/datastore/data-1/user-1
user-1@pbs has backup permission to this datastore
I added a datastore user-2 with the path:
/mnt/datastore/data-1/user-2
user-2@pbs has backup permission to this datastore
No one has permission to the main datastore data-1
Now we can use user-1@pbs and datastore user-1 in cluster 1
and user-2@pbs and datastore user-2 in cluster 2
The storage is for all, and there can be no conflicts of IDs.
Maybe someone find this usefull.
If you want to use a central PBS for more then one cluster, you run into the problem of double IDs.
Ok, you can use different datastores, but then you have to know the expected sizes.
We did an other aproach:
We build one big datastore and afterwards we add datastores via absolute backing path which are below the big datastore.
Datastore data-1 (the main ZFS datastore) is located in:
/mnt/datastore/data-1
So I added a datastore user-1 with the path:
/mnt/datastore/data-1/user-1
user-1@pbs has backup permission to this datastore
I added a datastore user-2 with the path:
/mnt/datastore/data-1/user-2
user-2@pbs has backup permission to this datastore
No one has permission to the main datastore data-1
Now we can use user-1@pbs and datastore user-1 in cluster 1
and user-2@pbs and datastore user-2 in cluster 2
The storage is for all, and there can be no conflicts of IDs.
Maybe someone find this usefull.