Hi,
I have an application that connects to DB2 8.1 server on Redhat Linux ES 3; it uses the DB2 8.1 CLI to talk to the DB2 server. This application serves as an intermediary (middleware) for client apps that connect to it via odbc.
|Client|-----ODBC----->|Application|----DB2 CLI---->|DB2|
When a client app exits, my application stops the thread that has been servicing the client. (I'm using the ACE framework). At that point, I see the following error message in db2diag.log:
------------------------------------BEGIN DB2DIAG.LOG SNIPPET----------------------------------------------
2004-09-09-18.34.23.415575 Instance:db2inst1 Node:000
PID:2994(db2agent (NAVYA)) TID:3021703520 Appid:*LOCAL.db2inst1.064700013347DRDA Communication Manager sqljcReceive Probe:30 Database:NAVYA
CCI Error:
0xBFFF64E0 : 0x00000047 G...
2004-09-09-18.34.23.431372 Instance:db2inst1 Node:000
PID:2994(db2agent (NAVYA)) TID:3021703520 Appid:*LOCAL.db2inst1.064700013347DRDA Communication Manager sqljcReceive Probe:30 Database:NAVYA
DIA0001E An internal error occurred. Report the following error code :
"ZRC=0x8136001C".
2004-09-09-18.34.23.454892 Instance:db2inst1 Node:000
PID:2994(db2agent (idle)) TID:3021703520 Appid:*LOCAL.db2inst1.064700013347
data protection sqlpterm Probe:3200
Database shutdown with next LSN of 000000000426800C.
------------------------------------------END DB2DIAG.LOG SNIPPET--------------------------------------------
Subsequent client connections to my application can no longer access db2. Interestingly enough, if I SIGKILL my application just after the client has recieved the results of the query, I get the same above message from DB2. So my client going away and SIGKILLING the application have the same effect on DB2.
My application does not encounter this problem on Solaris or AIX, so I am wondering if this may have to do with the multithreadedness (or lack thereof) of the DB2 8.1 CLI for Linux ?
Does this kind of error look familiar to anyone ?
Thanks for any advice,
Colm.