We have a TSM 5.4.2.0 server running under mainframe Linux. We run
reclamation of our copy storage pool with five concurrent processes.
The copy storage pool and the corresponding primary storage pool
have different device classes, with five tape drives available for
each device class. We will occasionally find one of the reclamation
processes reporting that it is waiting for a mount point when there
is in fact a pain of tape drives available for use. When this
happens, 'q process' reports no other processes using tapes,
'q session f=d' reports no sessions using tapes, 'q drive' and
'q path' show all the expected drives and paths defined and on-line,
'q mount' shows only four drives of each type in use, and 'q restore'
reports no active or restartable restores. When reclamation finally
ends and other types of processes start, the new processes have no
trouble using all of the available tape drives. The reclamation
process is ostensibly waiting for a mount point in the device class
used for copy pool tapes. However, when 'q process' reports
legitimate mount point waits for processes that use sequential
volumes from two different device classes, the device class is
frequently wrong.