Taking care of Fail Back in HDR

Again this is another surprise for me in Informix that failover is not automatic. In fact they expect the application logic to be written keeping replication in mind. How crappy is that ? For all the guys that use HDR how do you take care of the primary going down ? The manual mentioned 3 methods : Using DBPATH Changing Connectivity information Using INFORMIXSERVER. Which one is the preferred method ?

On Wed, 13 Aug 2003 08:36:49 -0400, Jack A wrote: There is another option. Define a group in the sqlhosts file, much as you would from ER, include both servers in the group with the primary server listed first. Make the 'e=' entry name the backup server. When the primary goes down any application just needs to reconnect when it detects the loss of connection, reprepare any statements and redeclare cursors and voila! You're back in business. Art S. Kagel > Again this is another surprise for me in Informix that failover is not > automatic. In fact they expect the application logic to be written > keeping replication in mind. How crappy is that ? > > For all the guys that use HDR how do you take care of the primary going > down ? The manual mentioned 3 methods : > > Using DBPATH > Changing Connectivity information > Using INFORMIXSERVER. > > Which one is the preferred method ?

On Wed, 13 Aug 2003 08:36:49 -0400, Jack A wrote: OH! Forgot to mention. You would set INFORMIXSERVER or the '@servername' clause in a CONNECT TO statement to point to the group rather than one of the replicants. Art S. Kagel > Again this is another surprise for me in Informix that failover is not > automatic. In fact they expect the application logic to be written > keeping replication in mind. How crappy is that ? > > For all the guys that use HDR how do you take care of the primary going > down ? The manual mentioned 3 methods : > > Using DBPATH > Changing Connectivity information > Using INFORMIXSERVER. > > Which one is the preferred method ?