Originally posted by: Robert_Willcox
We have a client that is getting sockets disconnected every 5 minutes. Weve noticed an APAR that indicates a bug in AIX 5.3 TL 4. The client is not willing to take their system down unless they know applying Service Pack 3 (w/ this fix) will correct the problem. The symptoms we see are that we have a TCP/IP socket connection, our host transmits a packet with the ACK sequence number set to X, our host then gets back a sequence number that is (X 1), this repeats every 30 seconds or so, for 4 minutes or so, until the host receives a RST and the socket connection is restarted for a while. This goes on and on. Our network engineers say that sequence numbers should never go/come in decreasing order and usually this symptom points to hardware failure. However the problem started when the client upgraded to AIX 5.3 TL 4.
There is a new APAR for 5.3 Service Pack (SP) 3, released last month that could be the problem here. We haven't installed it at our clients yet. But am bringing it up to converse w/ the experts. The APAR:
IY89429: AIX STOPS RESPONDING TO KEEPALIVES 06/11/01 PTF PECHANGE
A fix is available
Obtain fix for this APAR
APAR status
Closed as program error.
Error description
When a TCP connection goes idle, and remote end (need not
be AIX) sends TCP keepalives, AIX will respond to 5
keepalives and stop responding after that. However, it
will respond to FIN when closing the connection.
Local fix
Problem summary
***************************************************************
*USERS AFFECTED: *
-
All users with the following filesets at these levels *
-
bos.net.tcp.client 5.3.0.30
-
bos.net.tcp.client 5.3.0.31
-
bos.net.tcp.client 5.3.0.32
-
bos.net.tcp.client 5.3.0.40
-
bos.net.tcp.client 5.3.0.41
-
bos.net.tcp.client 5.3.0.42
-
bos.net.tcp.client 5.3.0.43
-
bos.net.tcp.client 5.3.0.44
-
bos.net.tcp.client 5.3.0.50
-
bos.net.tcp.client 5.3.0.51
-
bos.net.tcp.client 5.3.0.52
*************************************
#AIX-Forum