You may be running out of file handles. Based on your numbers, I'd guess there 
is a per-process limit of 1024 file handles.  Your storage volumes are using 
most of them and there aren't enough left over for TSM to do what it needs to 
(open a message file, etc).  

Assuming mainframe Linux is equivalent to other platforms, look into the ulimit 
command (ulimit -a will show the various resource limits). 

-Ken 


On Apr 7, 2011, at 9:30, "Thomas Denier" <Thomas.Denier AT JEFFERSONHOSPITAL 
DOT ORG> wrote:

> We have a TSM 5.5.4.0 server running under mainframe Linux. The server has
> two random access disk storage pools with a total of 726 volumes. Most of
> the volumes are 2 GB. A few are smaller to fit in space left over after
> populating file systems with as many 2 GB volumes as possible. Yesterday we
> attempted to add 233 more volumes to one of the pools. We had added 214
> when a volume formatting process failed. Shortly after that we starting
> seeing a wide range of errors. Reclamation processes failed. The server
> refused TCP/IP connection requests from both nodes and administrative
> command line clients. Sessions for scheduled backups (with prompt mode
> scheduling) hung. The activity log reported that message texts were
> unavailable for a variety of message numbers. Many header fields in query
> output contained something like 'HEADER NOT AVAILABLE'. The server was
> unable to write accounting records. I restarted the server, and the
> symptoms came back within minutes after the restart. I removed the new
> volumes and restarted the server again. The server then behaved normally.
> 
> Am I correct in suspecting that the problem had to do with the number
> of storage pool volumes, and that I will be able to enlarge the storage
> pool safely if I replace existing 2 GB volumes with volumes in the 10
> to 20 GB range, and use the same size for new volumes?