Restart Issue in Cluster environment

Hi,

We are using postrges 9.0.3 cluster environment, with replication patch.

In a normal condition, if I restart current active postgres in standby
mode and current standby postgres in active mode, we are facing one
blocking issue.

On restarting current standby postgres in active mode, the flag
internally maintained by postgres engine (select pg_is_in_recovery()) is
not getting changed. Because of this even though postgres is started
properly it is in read-only mode.

The restart of current standby postgres in active mode is valid because
both postgres are in sync and as long as both postgres are in sync I can
start any postgres in active mode.

So in such scenario, is there any option to forcefully change the flag?
Or postgres ideally should change this flag?

Regards,
Prakash