Hi Carel-Jan,

<snip>Now, we have a one-off requirement to switch roles between primary and the physical standby for couple of days.<snip>

Pardon me, but I think what Carel-Jan was trying to emphasize is that there is no such thing as a one-off requirement to switch between primary and standby roles.

If the ability to “fail-over” and “fail-back” is not routinely tested and practiced, then when an actual need arises you cannot be reasonably certain it will work, you will not know reasonable bounds on how long it will take, and you will be at risk to an increased window of accumulated software change and functionality since your last previous test. This leads to all manner of unfortunate adventures such as new client applications that were not integrated with the mechanism to point production connections to the “standby”.

In a real business continuation situation, the situation cannot be “I think I’d better do a test.”

Regards,

mwf