ldmwndletsm
ADSM.ORG Senior Member
- Joined
- Oct 30, 2019
- Messages
- 232
- Reaction score
- 5
- Points
- 0
After reading a lot about the MAXSCRATCH storage pool parameter, I just want to be clear on my understanding.
EXAMPLE: Storage pool A has collocation set to group, there are 1000 private volumes defined, MAXSRATCH is set to 1500, there are no empty predefined volumes, and there are 10 filling tapes. So the storage pool can grow to a maximum of 1500 volumes. If we reduced the MAXSCRATCH down to 1025 then the pool could grow by another 25 volumes to a max of 1025. If we reduced it down to 1000 then backups will have to restrict themselves to those 10 tapes regardless of how many scratch tapes exist or will in the near future (pending resusedelay). As long as there's at least 1 filling tape with nodes in that same collocation group then the backups will use that tape; otherwise, if there are no available tapes for that collocation group then it will select the tape with the lowest usage and bypass the collocation by group restriction, but backups will still run.
1. Is that correct ?
This was based on what I read in https://www.ibm.com/docs/en/tsm/7.1.0?topic=collocation-how-server-selects-volumes-enabled
2. And MAXSCRATCH is not the number of scratch tapes that can be used above and beyond the number of private volumes in the storage pool. Instead it's the max size in number of volumes that the storage pool can occupy, right ?
3. BUT if instead MAXSCRATCH was set to a value that's less than the number of volumes in use then this seems a bit of a contradiction based on the IBM documentation for "update storage pool". So in the above example, if MAXSCRATCH was changed to 900 then TSM can still work with the 1000 tapes already defined, not just 900 of the 1000, right ? And it can still use any filling tapes. So in theory, this number could be set to 1, and as long as you still had enough filling tapes then backups would still work, at least until you had no more filling tapes ?
4. What about "move data" ? Will it function similarly to a backup, as discussed in the above (1) IBM link ?
I was unable to find anything definitive on this in the documentation. So if there were no filling volumes that contained any nodes in that collocation group, and the number of private volumes <= MAXSCRATCH so TSM cannot use a scratch volume then would TSM simply move the data to one of the filling tapes with the most available space, as with backups, and likewise bypass any collocation by group restrictions ? Or would it fail altogether ?
EXAMPLE: Storage pool A has collocation set to group, there are 1000 private volumes defined, MAXSRATCH is set to 1500, there are no empty predefined volumes, and there are 10 filling tapes. So the storage pool can grow to a maximum of 1500 volumes. If we reduced the MAXSCRATCH down to 1025 then the pool could grow by another 25 volumes to a max of 1025. If we reduced it down to 1000 then backups will have to restrict themselves to those 10 tapes regardless of how many scratch tapes exist or will in the near future (pending resusedelay). As long as there's at least 1 filling tape with nodes in that same collocation group then the backups will use that tape; otherwise, if there are no available tapes for that collocation group then it will select the tape with the lowest usage and bypass the collocation by group restriction, but backups will still run.
1. Is that correct ?
This was based on what I read in https://www.ibm.com/docs/en/tsm/7.1.0?topic=collocation-how-server-selects-volumes-enabled
2. And MAXSCRATCH is not the number of scratch tapes that can be used above and beyond the number of private volumes in the storage pool. Instead it's the max size in number of volumes that the storage pool can occupy, right ?
3. BUT if instead MAXSCRATCH was set to a value that's less than the number of volumes in use then this seems a bit of a contradiction based on the IBM documentation for "update storage pool". So in the above example, if MAXSCRATCH was changed to 900 then TSM can still work with the 1000 tapes already defined, not just 900 of the 1000, right ? And it can still use any filling tapes. So in theory, this number could be set to 1, and as long as you still had enough filling tapes then backups would still work, at least until you had no more filling tapes ?
4. What about "move data" ? Will it function similarly to a backup, as discussed in the above (1) IBM link ?
I was unable to find anything definitive on this in the documentation. So if there were no filling volumes that contained any nodes in that collocation group, and the number of private volumes <= MAXSCRATCH so TSM cannot use a scratch volume then would TSM simply move the data to one of the filling tapes with the most available space, as with backups, and likewise bypass any collocation by group restrictions ? Or would it fail altogether ?
