Technical Tips

How to resolve a GC516 GOALSYS - termination due to CSA buffer shortage

This situation may occur when the number of CSA Buffers is too small to accommodate the number of concurrent requests from the cross-region
partner(s). Consider the following example:

• PHOENIX/PREFERENCE is running at a much lower priority than CICS
• A maximum of 100 cross-region users can concurrently access PHOENIX/PREFERENCE (XREGMAX=100)
• The number of LDAS= and # CSA Buffers may be default values (this would be 100/8 = 13 rounded)
• If multiple users request help, the first 13 will be assigned a CSA buffer.
• Until processing has completed for at least one of the first 13, new requests will be queued to retry for up to 90 seconds or when a CSA Buffer becomes available (which ever is shorter).
• If the system is busy and PHOENIX/PREFERENCE is not dispatched, then when the 90 seconds are up users receive a GC516.
• It is possible that some users are processed, but not all are processed within the 90 seconds waiting period due to the low priority of PHOENIX/PREFERENCE.

If the PHOENIX/PREFERENCE system is running at a priority equal to or higher than the cross-region partner(s) and both XREGBN= and LDAS= were not specified, then the default value for XREGBN= should be fine (maximum users/8).

If PHOENIX/PREFERENCE is running at a lower priority than its cross-region partner(s), then the default value for the XREGBN= parameter is too small. A value equal to the XREGMAX= parameter may be required to eliminate this termination condition. In the example above where XREGMAX=100, modify XREGBN so that XREGBN=100 instead of =8. This will not speed up the processing. It will only prevent the user's request from being terminated. To speed up the through put, the PHOENIX/PREFERENCE priority must be set equal to or higher than the cross-region partner(s).