JR47520: HPR HANG DUE TO HOST GAP MESSAGE PACKET BEING DROPPED IN THE NETWORK
A fix is available
Closed as program error.
A remote VTAM host has fixed pacing window size of 1. As the VTAM attempts to send data which needs more than 1 RU to complete it requests another pacing window to complete its transfer with a pacing indicator. CS/AIX sends an IPR with a status request (SR) with RASAP set. The remote VTAM host responds with a status with the Received Sequence Number (RSN) set to our BSN before the IPR, indicating that the remote host never received the IPR. However, the session continues to hang as CS/AIX never attempts to retransmit the IPR packet.
USERS AFFECTED: All users of HPR PROBLEM SUMMARY: Directly connected HPR/IP connection can hang up if packets are lost in both directions in quick succession and there is no other activity on the connection to drive retries. The gap detected packet from the mainframe was dropped in the network.
Code has been changed to restart the short request timer when a status packet is received which has an implicit gap detected, i.e. the partner system indicates that it has not seen the last system sequence number transmitted but has not reported the gap detected yet (it is waiting for the refifo timer to expire).
Reported component name
COMM SERV NT 6.
Reported component ID
Last modified date
APAR is sysrouted FROM one or more of the following:
APAR is sysrouted TO one or more of the following:
Fixed component name
COMM SERV NT 6.
Fixed component ID
Applicable component levels