Info Letter: Potential problem where TS7700 running code level R2.1 PGA7, any R3.0, any R3.1 or R3.2 GA with TPF workload may result in system reboot and impact to customer operation

Answer

Background:
TS7700 code level R2.1 PGA7 (8.21.0.178) introduced a new optional advanced function for tuning performance for specific TPF workloads by adding additional device buffering space.
With this optional feature enabled for a logical device, additional buffer space is allocated when a tape is loaded on the device by TPF.
This additional device buffering should have been disabled by default, yet due to accidental enablement it consumes an additional 100 MB of memory per device used by TPF.
The unintended enablement may cause the TS7700 to run out of memory and reboot and thus impact customer operation.

In addition a problem was fixed where if a TPF job is run on a device and then a non-TPF job (e.g. MVS) is run on that same device, this may result in a job failure or even a TS7700 reboot that affects all jobs on the cluster.

Solution (Procedure):
This problem is fixed in R3.2 PGA1 (8.32.1.8) and above.
Any V07/VEB running any R3.0, any R3.1 or R3.2 GA level of code used for TPF workload is recommended to upgrade to code level R3.2 PGA1 or above.
For recommended levels check here: http://www-01.ibm.com/support/docview.wss?uid=ssg1S1010472
For V06/VEA running R2.1 PGA7 level of code used for TPF workload vtd_exec.174 providing mitigation has been made available. In addition a manual procedure is available.