Steve White wrote:
>
> hi-all,
>
> i've a question regarding tape reclaimation.  how do i find out if tapes
> are getting returned to the scratch pool when adsm is done with them?
> are copypool/offsite empty tapes ever returned to scratch pool?

ADSM will create a volume history record of type STGDELETE when it
returns a volume to scratch status. You can use the 'query volhist'
command to find out whether these volume history records are being
created. If volumes are being returned to scratch status, they should
stop being visible to 'query volume' commands. If you are using an
automated library the 'query libvolume' command can be used to find
out explicitly which volumes are scratch volumes.

Empty volumes that were taken from the scratch pool and have an
access mode of 'offsite' will be returned to scratch status when
the access mode is changed to something else.

> here are the admin sched commands i currently run:
> (reusedelay is 1 day)
>
> 1:00 pm :
> UPDATE STGPOOL COPYPOOL RECLAIM=75  (i also reset these both to 100)
> UPDATE STGPOOL TAPEPOOL RECLAIM=75
>
> 5:00 pm :
> UPDATE VOLUME * ACCESS=OFFSITE
>  WHERESTG=COPYPOOL
>  WHEREACCESS=READWRITE,READONLY
> 5:00 pm :
> UPDATE VOLUME * ACCESS=READWRITE
>  WHERESTG=COPYPOOL WHEREACCESS=OFFSITE
>  WHERESTATUS=EMPTY,PENDING

Several features of this pair of update commands would make me
nervous. It appears that volumes with a status of 'pending' are
considered eligable for retrieval from the vault. The pending status
is supposed to mean that the volumes might be needed for use in
conjunction with an aging backup of the ADSM database. If this is
true, the volumes with that status should probably stay at the
vault. It also appears that vaulted volumes are considered available
for reuse as soon as they become empty. My site's policy is to
declare such volumes available for reuse only when they actually
arrive at the machine room. The pair of commands has the net
effect of resetting an access mode of 'readonly' to 'readwrite'.
Doing this as a blanket policy runs a real risk of suffering
the same tape errors repeatedly.

> are there other admin commands regarding reclaimation i should be
> running?

As noted above, a volume history record of the STGDELETE type is created
each time a tape is returned to scratch status. In addition, a volume
history record of the STGNEW type is created each time a scratch volume
is put into service as a storage pool volume. You will probably want to
run 'delete volhistory' commands regularly to prune older volume history
records of these types.

> i guess im not sure i understand the reclaim process.  Do tapes ever get
> returned to the scratch pile?
>
> could someone explain? please.
>
> thanks,
> steve white
>