HADR is the preferred solution for HA and DR on DB2 LUW. It has the following benefits:
Universal solution
HADR is integrated in DB2 LUW. It is available wherever DB2 is.
TCP is used for primary and standby connection. No special hardware or software is needed.
No limitation on distance between primary and standby.
Deploy standby locally in the same room for HA, or remotely across the continent for DR
Deploy both local and remote standbys for HA and DR at the same time, using
multiple standby
(starting in DB2 V10.1).
Ease of use
Setup
only requires a few configuration parameters (primary/standby network address, and hadr_syncmode)
Whole database replication. All logged activities in database are replicated (DDL and DML).
Only 3 commands to manage: start/stop/takeover HADR
One command does it all for failover (or role switch). Just say "takeover HADR"
Choice of protection levels
Choose the best
sync mode
for your business need: sync, nearSync, async, superAsync
Zero data loss protection available via patented IBM technology
peer window
.
Performance
Physical replay for best performance. Standby knows exactly which page, which row slot to update. Replay is simple and fast. Standby is exact clone of the primary.
Parallel replay on the standby. It uses the same mechanism as high performance parallel recovery.
Automation
Automated failover supported by integrated cluster manager.
Automatic client reroute after failover or role switch.
Automatic reconfiguration after failover or role switch in multiple standby. All standbys are auto reconfigured to connect to the new primary.