Tips for Replication

[svk]

Hello everyone and especially those who have worked on/implemented a snapshot replication :

Can you please highlight any issues/concerns/problems you have encountered and need special mention/attention ?

Can someone outline a scenario in which the snapshot database crashed, incomplete recovery had to be made and now the question is re-synching the snapshots with master tables. Is re-creating the snapshots the only solution ?

Anything that you feel important about such a replication env, is invaluable input for me....

[jm]

Here are some of my notes:

master site:
1. job queue interval and refresh job interval, if it too small/frequent, it will create "hugh" volumn of arch log files.
2. using dedicated tablespace for snapshot log file

snapshot site:
1. create a dedicated tablespace for index
2. since snapshot only copy the primary key, no other index will be created. consider create/add index besides primary key by yourself.
3. you might need ALTER schema owner DEFAULT TABLESPACE to be the tablespace which the the primary index located -- if you want primary index created in "the" tablespace.
4. what if the refresh job "broken" due to other reason
e.g. network down, the refresh will stop after "16" failure.
you might need some "watchdog" to enable the job.

re-synching the snapshot??? I have the same question