IBM DB2 LAN-free backup fails with return code -72

Problem(Abstract)
After the backup runs for some time, the DB2 lanfree backup failed with SQL2033N and return code -72.

Symptom

The following messages are displayed in the db2diag.log file:

2014-10-27-21.07.19.678712+330 E329353406E496 LEVEL: Info
PID : 23967 TID : 140713299797760 PROC : db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB : TESTDB
APPHDL : 0-59736 APPID: *LOCAL.db2inst1.141027153704
AUTHID : DB2INST1 HOSTNAME: TSMTEST
EDUID : 969 EDUNAME: db2agent (TESTDB) 0
FUNCTION: DB2 UDB, database utilities, sqlubSetupJobControl, probe:1850
MESSAGE : Starting an online db backup.
...
2014-10-27-23.37.55.528047+330 E406903852E389 LEVEL: Error
PID : 29134 TID : 140737012004640 PROC : db2vend (db2med - 2637 (TESTDB)
INSTANCE: db2inst1 NODE : 000
HOSTNAME: TSMTEST
FUNCTION: DB2 UDB, database utilities, sqluvend, probe:1134
DATA #1 : TSM RC, PD_DB2_TYPE_TSM_RC, 4 bytes
TSM RC=0xFFFFFFB8=-72 -- see TSM API Reference for meaning.

2014-10-27-23.37.56.546906+330 E406904242E530 LEVEL: Error
PID : 23967 TID : 140712884561664 PROC : db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB : TESTDB
APPHDL : 0-59736 APPID: *LOCAL.db2inst1.141027153704
AUTHID : DB2INST1 HOSTNAME: TSMTEST
EDUID : 2637 EDUNAME: db2med.969.0 (TESTDB) 0
FUNCTION: DB2 UDB, database utilities, sqluMCWriteToDevice, probe:1015
MESSAGE : Media controller -- problem during writing to vendor device

2014-10-27-23.37.56.547549+330 E406904773E596 LEVEL: Error
PID : 23967 TID : 140712884561664 PROC : db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB : TESTDB
APPHDL : 0-59736 APPID: *LOCAL.db2inst1.141027153704
AUTHID : DB2INST1 HOSTNAME: TSMTEST
EDUID : 2637 EDUNAME: db2med.969.0 (TESTDB) 0
FUNCTION: DB2 UDB, database utilities, sqluMCWriteToDevice, probe:1015
MESSAGE : SQL2033N An error occurred while accessing TSM during the processing
of a database utility. TSM reason code: "".

2014-10-27-23.38.00.076366+330 E406973545E475 LEVEL: Severe
PID : 23967 TID : 140713299797760 PROC : db2sysc 0
INSTANCE: db2inst1 NODE : 000 DB : TESTDB
APPHDL : 0-59736 APPID: *LOCAL.db2inst1.141027153704
AUTHID : DB2INST1 HOSTNAME: TSMTEST
EDUID : 969 EDUNAME: db2agent (TESTDB) 0
FUNCTION: DB2 UDB, database utilities, sqlubcka, probe:1050
MESSAGE : Backup terminated.

Cause
The ResourceTimeout and IdleTimeout parameters are set to large values, for example, 10 hours. The interval at which the storage agent sends these keep-alive verbs to the server is 2.5 hours (10 hours / 4 = 2.5). For a detailed explanation about these parameters and options is available in Lanfree backup/restore operations fail due to firewall timeout. The reason the return code -72 is returned is that the client session is ended. It ends because either the firewall times out or the network session closes.

Diagnosing the problem
Tivoli Storage Manager activity log shows the session abort when DB2 backup fails with -72:
10/27/2014 21:06:11 ANR0406I (Session: 3000, Origin:TSMTEST_STGAGENT) Session 6 started for node TSMTEST_NODE (DB2/LINUXX8664) (Tcp/Ip 10.1.1.0(46029)). (SESSION: 3000)
...
10/27/2014 23:36:23 ANR0530W (Session: 3000, Origin:
TSMTEST_STGAGENT) Transaction failed for session 6 for node TSMTEST_NODE (DB2/LINUXX8664) - internal server error detected. (SESSION: 3000)

10/27/2014 23:36:43 ANR0480W (Session: 3000, Origin: TSMTEST_STGAGENT) Session 6 for node TSMTEST_NODE (DB2/LINUXX8664) terminated - connection with client severed. (SESSION: 3000)

Resolving the problem
Reduce the value of either the IdleTimeout or ResourceTimeout parameters on both the Storage Agent and Tivoli Storage Manager. The interval to send keep-alive verb frequently needs to be enough to prevent the timeout from occurring. For example, set ResourceTimeout to 60 minutes by entering "setopt RESOURCETIMEOUT 60". In this example, the interval to send keep-alive verb is every 15 minutes.