HADR - AIX

Hi,

If the HADR state is 'Disconnected' and commit transactions to the
Primary database, in the event of a Failover to the Standby database,
how do we determine if it is safe i.e. how do we know how far behind
the standby database is?

Basically what is the procedure to determine if it's safe to do a
failover to the standby with a takeover hadr .....force.

re: HADR - AIX

<agentlease@hotmail.comwrote in message
news:0481bdf8-eb8a-4453-9235-e294c03c17e7@x16g2000prn.googlegroups.com...It is only safe to do a takeover by force if the primary server is not
available to the application (crashed, network connection is not accessible,
etc). If you are automating the takeover with a Cluster Manager, then make
sure your instance is not on autostart.

You can run a command on the primary and standby to see where the log
position is. Normally if they are in peer state you can run it from either
server, but if in disconnected state you would have run it on each server to
get the log position of each. You have two choices:

db2 get snapshot for database on <dbname(there is section on HADR in the
output)
db2pd -d <dbname-hadr

re: HADR - AIX

You may also want to look at the HADR Peer Window feature, which can
help ensure that transactions do not commit on the primary without
replication to the standby, for a configurable period of time. This
is useful where there is a high desire for consistency at failover,
and esp. where failover is automated. Note that this choice for
consistency does trade off against availability, as when the HADR pair
is disconnected, the primary (in SYNC or NEARSYNC modes) cannot commit
transactions during Peer state, nor in the new Disconnected Peer state
that comes with the Peer Window feature.

Some intro material here:

http://publib.boulder.ibm.com/infoce.../c0051997.html

Regards,
- Steve P.
--
Steve Pearson, DB2 for Linux, UNIX, and Windows, IBM Software Group
"Portland" Development Team, IBM Beaverton Lab, Beaverton, OR, USA