Standby database Transport Lag

Oracle 10.2.0.3 EE with Data Guard Physical Standby configured with Asynchronous transport and real time apply. Platform RHEL 4 x86_64.

How can I determine the exact lag between when a transaction is committed on the Primary database and when that transaction gets written into the Standby Redo Log on the physical standby..?

OEM Grid has a metric called Transport Lag which seems to be related to this time but it always reports a Transport Lag of between 5 and 15 seconds even when our redo generation rate is very low (200 kbits/ sec). The primary and standby are connected via a relatively underutilised  gigabit network so this Lag is unnecessarily high.

So, firstly – is the Transport Lag statistic accurate (which comes from select * from v$dataguard_stats where name = ‘transport lag’;)

If not how can I accurately calculate the difference between the time a transaction commits on the primary and the time that transaction arrives on the Standby (written to the Standby Redo Log)..?

In the Oracle published doc “Data Guard Redo Transport Network Best Practices 10gR2’ the author states that he has a lag of less than 2 seconds with a redo generation rate of 2Mbytes/sec on a network with a latency of less than 1ms. Our network latencies are also less than 1ms and our redo generation rate is 1 tenth of theirs yet our Transport Lag is consistently between 5 and 15 seconds.

Any ideas..?

Thanks Received on Thu Nov 27 2008 - 18:04:59 CST