We have recently implemented an Attunity replication between 2 servers (From Oracle -> to MSSQL).
Sometimes, we are facing high latency / incoming changes between the 2 servers.
and most of the time we can observe (on SQL Server) a query from attunity program, on a specific table, blocking users queries :
When the query complete, there is no more latency between servers.
According to sp_woisactive, the query is in RUNNABLE state. Trying to start a select on the lock table, my query is blocked, waiting for the attunity query to end.
If we stop attunity replication, the query end and all other transactions runs fines.
Have you Any ideas why this query is taking so much time ? How can I diagnostic this ? it's just a delete on a specific id. Moreover, Indexes are all up to date and there is no fragmentation.