FW: SR 603066805 : FW - Firewall in transparent mode with excessive packet drops

JKCGI

Description of the DB2 Connect problem:

Users of two LTC client-server applications are experiencing intermittent
disconnects from the mainframe. Users of the Retail Workflow application
and the Group LTC Care application are experience SQL errors after
attempting to execute a transaction after leaving the application idle for
approx. 30 minutes.

The WLTC application architecture was "cloned" from GLTC Care. The
architecture is as follows:

PowerBuilder 9.0 GUI running on an NT client
DB2 Connect 8.1 gateway for connectivity to the mainframe
DB2 on MVS database
DB2 Stored procedure using MVS Cobol modules in composite link for the
business logic
Communication uses ODBC ---> DB2 CLI ---> TCP/IP ---> DB2 Connect
gateway ---> MVS DB2 on the mainframe

The SQL error messages that are being displayed are as follows:

FYI , I just saw the messages in DB2 :

DB2TMSTR

21.01.55 STC00397 DSNL511I - DSNLIENO TCP/IP CONVERSATION FAILED 625

625 TO LOCATION 165.71.112.68

625 IPADDR=165.71.112.68 PORT=1593

625 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000
21.11.58 STC00397 DSNL511I - DSNLIENO TCP/IP CONVERSATION FAILED 864

864 TO LOCATION 165.71.112.67

864 IPADDR=165.71.112.67 PORT=2650

864 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000
21.45.27 STC00397 DSNL511I - DSNLIENO TCP/IP CONVERSATION FAILED 185

185 TO LOCATION 165.71.112.68

185 IPADDR=165.71.112.68 PORT=1572

185 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000

DB21MSTR

16.45.41 STC04970 DSNL511I @ DSNLIENO TCP/IP CONVERSATION FAILED 769

769 TO LOCATION 165.71.112.68

769 IPADDR=165.71.112.68 PORT=4795

769 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000
18.52.59 STC04970 DSNL511I @ DSNLIENO TCP/IP CONVERSATION FAILED 056

056 TO LOCATION 165.71.112.67

056 IPADDR=165.71.112.67 PORT=2283

056 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000
20.56.54 STC04970 DSNL511I @ DSNLIENO TCP/IP CONVERSATION FAILED 166

166 TO LOCATION 165.71.112.68

166 IPADDR=165.71.112.68 PORT=1582

166 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000
21.37.06 STC04970 DSNL511I @ DSNLIENO TCP/IP CONVERSATION FAILED 929

929 TO LOCATION 165.71.112.68

929 IPADDR=165.71.112.68 PORT=1571

929 SOCKET=RECV RETURN CODE=1127 REASON
CODE=00000000
Just to document the test and result:

PC #1 - Logged into DB2T (db WLTCDBD20) through application WLTC under
HG00440 and ran transaction.
Also logged into DB21 (db WLTDBP10) via PowerBuilder under
HG00440 and ran a query different than above transaction.

Test #1: Waited :45 minutes after logging in and re-ran the same
transaction in WLTC as initially done and the same query in PoweBuilder as
initially done. This worked without any problem.

Ran first transaction at 6:30pm. Ran transaction a second time at 7:15pm.
Test passed. (session was idle for 45 minutes)

Ran PB query at 6:30pm and again at 7:15pm. Test passed.(session was idle
for 45 minutes)
Test #2: Waited until 8:30pm and ran WLTC transaction again. Transaction
failed displaying error message below. (session was idle for
75 minutes - 7:15pm to 8:30pm)

Ran PB query again at 8:30. Query failed with same
message.(session was idle for 75 minutes - 7:15pm to 8:30pm)

PC #2:
Test #1
Logged into WLTC under WLTCNB1 into DB2T (db WLTCDBD20) at 6:47 pm and ran
same transaction as above.
Logged into WLTC under HG00440 into DB21 (db WLTDBP10) at 6:47pm and ran
same transaction as above.
Test#2
Tried WLTC transaction in DB2T at 8:53pm and transaction failed. Same
message as PC#1.(session was idle for 126 minutes)

Tried WLTC transaction in DB21 again at 9:11pm and transaction failed.
Same message as PC#1. (session was idle for 144 minutes)

ANy help would be great!

Feb 26 '06 #1

Gert van der Kooij

In article <c6c28b7a1b8042ffdb385bf67f8234f9
@localhost.talkaboutdatabases.com>, JKCGI (jf****@comcast.net)
says...

Description of the DB2 Connect problem:

Users of two LTC client-server applications are experiencing intermittent
disconnects from the mainframe. Users of the Retail Workflow application
and the Group LTC Care application are experience SQL errors after
attempting to execute a transaction after leaving the application idle for
approx. 30 minutes.

The WLTC application architecture was "cloned" from GLTC Care. The
architecture is as follows:

PowerBuilder 9.0 GUI running on an NT client
DB2 Connect 8.1 gateway for connectivity to the mainframe
DB2 on MVS database
DB2 Stored procedure using MVS Cobol modules in composite link for the
business logic
Communication uses ODBC ---> DB2 CLI ---> TCP/IP ---> DB2 Connect
gateway ---> MVS DB2 on the mainframe

The TCP error 1127 as displayed in your messages means a connection
timeout error. This looks like a TCP/IP configuration problem. I'm
not a TCP/IP expert but the Keep-alive interval might be worth
checking.

Feb 26 '06 #2