This problem occurs during initialization of HADR in pureScale,
when
the HADR_TARGET_LIST database configuration is not set
correctly.
In particular, the HADR_TARGET_LIST for the primary database is
set to the hosts in the pureScale cluster for the primary,
instead of
the hosts in the cluster for the HADR standby database.
With such incorrect configuration, the START HADR AS PRIMARY
command could fail with SQL1776N reason code 6. The following
messages can be found in db2diag.log:
2021-01-15-03.31.58.413399-360 I1642734894A496 LEVEL: Info
PID : 29164230 TID : 209500 PROC : db2sysc 2
INSTANCE: db2inst1 NODE : 002 DB : MYDB
HOSTNAME: myhost2.somedomain.com
EDUID : 209500 EDUNAME: db2hadrp.2.1 (MYDB) 2
FUNCTION: DB2 UDB, High Availability Disaster Recovery,
hdrHandleHsAck, probe:43900
DATA #1 :
Handshake HDR_MSG_REDIRECT message is received from
myhost1:59000 (10.50.31.101:59000)
2021-01-15-03.31.58.415990-360 I1642736389A689 LEVEL: Warning
PID : 29164230 TID : 209500 PROC : db2sysc 2
INSTANCE: db2inst1 NODE : 002 DB : MYDB
HOSTNAME: myhost2.somedomain.com
EDUID : 209500 EDUNAME: db2hadrp.2.1 (MYDB) 2
FUNCTION: DB2 UDB, High Availability Disaster Recovery,
hdrHandleRedirectionHandshake, probe:16110
MESSAGE : ZRC=0x878001AF=-2021654097=HDR_ZRC_EVIL_TWIN
"The database is configured as a primary database, but another
primary database already exists."
DATA #1 :
A split-brain has been detected. Sending a redirect message to
the standby so that it can shut down.
Db2 should have detected the incorrect configuration and prevent
the above actions to be taken as a result. |
