Here is how you can change the priorities in ATRESYS vs. in which tape pool the tape ends-up:
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager ls -v L94261
Tape Status Media Pool Bytes Usage Last Written Mode Priority Operator
L94261 1/6 Entered 2025-09-03 22:43 LTO9 backup_ZVAULT_1 688.0G 4% 2025-05-03 22:57 auto low tape-local
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager change -p high -v L94261
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager ls -v L94261
Tape Status Media Pool Bytes Usage Last Written Mode Priority Operator
L94261 1/6 Entered 2025-09-03 22:43 LTO9 backup_ZVAULT_1 688.0G 4% 2025-05-03 22:57 auto high-tolabel tape-local
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager change -p low -v L94261
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager ls -v L94261
Tape Status Media Pool Bytes Usage Last Written Mode Priority Operator
L94261 1/6 Entered 2025-09-03 22:43 LTO9 backup_ZVAULT_1 688.0G 4% 2025-05-03 22:57 auto low tape-local
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager change -p low -d erase -v L94261
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager ls -v L94261
Tape Status Media Pool Bytes Usage Last Written Mode Priority Operator
L94261 1/6 Entered 2025-09-03 22:43 LTO9 backup_ZVAULT_1 688.0G 4% 2025-05-03 22:57 auto low-erase tape-local
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager change -p high -d erase -v L94261
[tape-local@ctaproductionfrontend11 ~]$ cta-ops-repack-manager ls -v L94261
Tape Status Media Pool Bytes Usage Last Written Mode Priority Operator
L94261 1/6 Entered 2025-09-03 22:43 LTO9 backup_ZVAULT_1 688.0G 4% 2025-05-03 22:57 auto high tape-local
I think what you need to do is just to change the /etc/cta-ops/cta-ops-config.yaml configuration file.
To clarify:
tolabel tape pool = tape will be labelled again and is immediately ready to be used
erase tape pool = tape had an I/O error and needs to be completely tested with cta-ops-admin tape mediacheck (destructive) to see if the tape can be reused at all (i.e. if the error is permanent or not)
Hi Jeff,
To give additional context: There are two different notions of priority here, which I guess is the source of confusion:
There is the notion of priority in CTA, which there is referred to by the name mount_policy.
There is the notion of priority within ATRESYS, which is used to determine when ATRESYS submits a scheduled repack from within its own database to CTA.
The ATRESYS priority is captured by an integer, as shown in the config exaple, where a higher number is higher priority.
The number is used for ATRESYS’s scheduling, and to map its notion of priority to the mount policies and destination pools configured in CTA.
Now, the problem this part of the config/code is trying to solve may be a bit CERN specific. We are trying to present the operator with a simple set of priority options:
low: “This is a run of the mill repack, do it whenever we get time”
high: “This is an important repack, probably a tape someone tried to retrieve data from, but the tape is now broken.”
But, we also want to distinguish slightly in terms of priority with regards to the destination pools Vlado mentions. Tapes that are to be reused (relabeled) have slightly lower priority than the ones which are to be erased (and removed from the system).
The idea with the config file was precisely to allow existing and new preferences to be configured, as people may be using different mount_policies. So in theory, if you just want to create new mappings between the high/low and destination pools then the config file should be sufficient.
The above was the solution that was good enough for our needs at the time. But if you figure out something neat then please consider not just a fork but a merge request to bring it back in.
Cheers,