Setting up database mirroring with sql anywhere 16.

[Former Member]

BH Hi, we are looking into database mirroring with sql anywhere 16. Out of the 3 computers (srv1 srv2, arbiter), which one should to be the most powerful? Also, if srv1 goes offline, then the arbiter automatically makes srv2 as the primary? Or does a network administrator has do something? Thanks so much for your help. Aron

[VolkerBarth · Accepted solution]

Out of the 3 computers (srv1 srv2, arbitrator), which one should to be the most powerful?

Usually srv1 and srv2 should have similar resources, as each one of them might be used as primary database. If you like to prefer one of srv1 or srv2 as primary (because that machine is more performant, you can set that explicitly with an according CREATE MIRROR SERVER clause. - Generally the primary server will have to do more work as it has to handle the normal client requests. However, you can choose to make the mirror server available for read-only queries, and in that case it's difficult to tell which server role would require more resources without knowing your particular workload...

The arbiter will require less resources as it does not fulfill real database server tasks - it just helps the other servers to find out if at least two of the three can communicate with each other. (Note that the arbiter can additionally be used as a normal database server for a different database.)

Also, if srv1 goes offline, then the arbitrator automatically makes srv2 as the primary?

That depends on the "database mirroring modes": For the default "synchronous mode", a failover can be done automatically - except in rare cases. For the other modes, a failover won't happen automatically unless you have set the auto_failover option.