Friday, October 30, 2015

Goldengate Replicat's blocking itself

Scenario: Multiple replicats working on same table and blocking itself 
Replicats are not moving, creating 14 hours lag. Data is out of sync between instances 


Root-cause
we are replicating very large volume tables and where we split replicat 1 into 6  using range function to avoid lag on replicat. so here, all 6 replicats working on same table. i.e 6 replicats will perform concurrent operation on the block. current operations on the block is controlled by storage parameter on database INIT Trans. on our database it is only 1. It is not allowing replicats to perform concurrent operations on the block resulting in 14 - 17 hour lag. 

Potential fix: 
Finally we increased storage parameter INIT trans parameter for the table from 1 to 10 
where replicat can perform concurrent operations on the table, which resolved this issue 

I strongly recommend, when we are splitting replicats to handle or replicat large volume data, I recommend to look into storage level parameter " INIT Trans " for the table. If the init trans parameter is n on the table. I recommend to split  single replicat into n+1 


---Nikhil Tatineni
--Oracle in memory 

Querys to monitor RAC

following few  Query's will help to find out culprits-  Query to check long running transaction from last 8 hours  Col Sid Fo...