Business Continuity and High Availability seem to be the at the top of customer priority lists at the moment. I previously blogged about the ability to use failover options on JDBC connections from within WAS. Continuing on from this I thought I would capture the high level points for setting up DB2 HADR for your CM8 or FileNet ECM system.
High Level overview of the main steps required to enable DB/2 HADR:
- enabled databases for archival logging or circular with logretain = ON
- update o/s servcies files to contain the DB2 HADR Services eg, db2_hadr_01 …. port50011
- set HADR db cfg parameters as below > db2 update db cfg for filenet using hadr_local_host SERVERB
> db2 update db cfg for filenet using hadr_remote_host SERVERA
> db2 update db cfg for filenet using hadr_local_svc DB2_HADR_24
> db2 update db cfg for filenet using hadr_remote_svc DB2_HADR_23
> db2 update db cfg for filenet using hadr_remote_inst DB2
> db2 update db cfg for filenet using hadr_timeout 120
> db2 update db cfg for filenet using hadr_syncmode SYNC
> db2 update db cfg for filenet using logretain on
> db2 update db cfg for filenet using LOGINDEXBUILD on
- Set the alternate server in the db directory as follows:
> db2 update alternate server for database filenet using hostname SERVERA port 50000
- Make a backup copy of the database
- Copy to standby server
- Restore backup as follows:
> db2 restore database filenet from e:\db2backups replace history file
- Update the db cfg by reversing the foregoing update db cfg commands so the local service is now standby and remote service is primarystart standby first
-
NB: Standby database should be roll forward state of pending before being able to activate hadr
> db2 rollforward db iccdb query status
- Activate the restored database:
> db2 activate database filenet
- Now you are ready to enable HADR on the standby – always do the standby first
> db2 start hadr on db filenet as standby - Now enable HADR on the primary as follows:
> db2 start hadr on db filenet as primary
- Useful commands to check the status are:
> db2 get snapshot for db on filenet | grep Role
> db2 get snapshot for db on filenet | grep State
- You can stop HADR using the following command on either server
> db2 stop hadr on db filenet
- To initiate a takeover of the primary from the standby issue the following:
> db2 takeover hadr on db iccdb
> db2 takeover hadr on db iccdb by force (Force failover of Automatic reroute client becareful as requires set up again to reactivate primary)
