ammu440,
Please read the doc marclant posted above regarding how to update the container-copy pools.
From the output of your message above his, it seems your Reclamation Threshold for your container-copy pool is set at 100. Also, appears that when you do run a reclaim, you will not see those volumes return to scratch until three days later.
The settings I'm looking at are:
Reclamation Threshold: 100
Offsite Reclamation Limit:
Maximum Scratch Volumes Allowed: 35
Number of Scratch Volumes Used: 33
Delay Period for Volume Reuse: 3 Day(s)
So, the value 100 says that volumes in this storage pool are not reclaimed. The general rule of thumb is to have this setting higher than 50, and at or lower than 85. This way you get two tapes for every one filled with the reclaim process. Again, it will take 3 days for those volume to return to scratch tapes based on the 'Delay Period for Volume Reuse' setting. The 'best practice' is not to return tapes to scratch until your oldest database backup is deleted. Just in case you need to restore your database back to a point in time. I know, that will go into a deeper discussion, but it is something you must take into consideration. I understand that you may wish to have scratch tapes right away, so you should update that value to whatever you feel comfortable with (Value of 0 will return to scratch almost immediately). Just remember if you have to change it, set it back to what your db2 database/DRM retention is.
Here's a little query that will help you determine what volumes you could reclaim based on the pct_reclaim. You will need to modify it to fit your environment and I will point you where
select volume_name,pct_reclaim from volumes where scratch='YES' and stgpool_name='PROTECTPOOL' and volume_name like '%VOLNAMELIKEHERE' and pct_reclaim>VALUEHERE order by pct_reclaim desc
So you will need to replace VOLNAMELIKEHERE with how your tapes are labeled (Hopefully with a common naming format, in my environment every tape has the type appended to it, like xxxxxL6 = LTO6, so I can use the statement like '%L6' ) and VALUEHERE by a percentage, say 85 for example, so you will see below pct_reclaim>85 . I went ahead and filled in your stgpool_name with the name listed above. If different, then you will need to modify 'PROTECTPOOL' as well.
In my environment, this is what I run and results:
Code:
select volume_name, pct_reclaim from volumes where scratch='YES' and stgpool_name='STGPOOLNAME' and volume_name like '%L6' and pct_reclaim>85 order by pct_reclaim desc
VOLUME_NAME: xxxxx1L6
PCT_RECLAIM: 100.0
VOLUME_NAME: xxxxx2L6
PCT_RECLAIM: 89.5
VOLUME_NAME: xxxxx3L6
PCT_RECLAIM: 89.1
So, those volumes above are 100% reclaimable, and 89% reclaimable or 11% used depending on how you look at it.
You will need at least one scratch tape free to reclaim that data. Here's a query for scratch tapes:
SELECT library_name,COUNT(*) FROM libvolumes WHERE status='Scratch' and volume_name like '%L6' GROUP BY library_name
Again, replace L6 with your common tape naming convention.
My ask if you could:
- Please let us know how many tape drives you have.
- Output of the query I provided with modifications to fit your environment, and with the reclaim_pct>51 so we may see what we are working with.
- Output of scratch volume query with modifications to fit your environment, so we know how many scratch tapes you have.
My concern is if you try to run a reclaim at 51%, and only have a few scratch tapes, and all tapes are half used; Your reclaim process could run a LONG time then run out of scratch tapes. That would affect your daily protect process. Chain of unintended consequences is what I'm looking to avoid.
My recommendation would be start small and slowly lower the reclaim threshold. As you can see, I chose 85, and it only had 3 tapes. If I ran the reclaim process at say 75 then that's nearly 40 tapes for me! Yes, there are other ways to limit the amount of tapes to be reclaimed with each protect process, but I'm trying to keep it simple for now.
Marclant,
I'm a bit fuzzy on this however, in the off chance ammu440 only has one tape drive, the reclaim process for container-copy's don't read data from tape right? its just reseeding data from the disk directory-container pool, so we can get away with a single tape drive?
Hope I didn't leave anything out, and hope this helps!