When customer sets up DB2 HADR along with TSA for automatic
failovers, then failover may not happen when DB2NODE is set.
db2diag.log message may look like as follows:
2012-07-20-10.17.00.431628-240 I61764A283 LEVEL: Info
PID : 6947032 TID : 1 PROC : db2gcf
INSTANCE: db2inst1 NODE : 000
EDUID : 1
FUNCTION: DB2 Common, Generic Control Facility, gcf_start,
probe:230
DATA #1 : String, 8 bytes
Unknown
The above return code "UNKNOWN" means that there was an error
when trying to get the state of hadr which causes the failover
to fail.
In syslogs, hadrv97_start.sh is being issues and returns 0
return code which means success, however, hadrv97_monitor.sh
returns error code 2 which means offline:
Jul 19 10:53:45 HOST user:notice
/usr/sbin/rsct/sapolicies/db2/hadrV97_start.ksh[18481234]:
Returning 0 : db2inst1 db2inst1 sample
Jul 19 10:53:46 HOST user:notice
/usr/sbin/rsct/sapolicies/db2/hadrV97_monitor.ksh[10485938]:
Returning 2 : db2inst1 db2inst1 sample
In order to confirm this issue, run the following command as db2
instance owner:
<instance_home>/sqllib/bin/db2egcf -i <instance_name> -i
<instance_name> -h <dbname> -s
The output should be "UNKNOWN"
This is because DB2NODE db2 variable is set to 0 on non DPF
environment.
Once the variable is unset, the correct state is returned and
failover happens successfully. |
