Solved: Is is possible to stay failed over?

ciscocsoc · ‎08-14-2008

Hi,

With ACE 2.1.0a is it possible to arrange for a backup serverfarm to remain the active serverfarm even when the primary serverfarm recovers?

I have a user who wants a SQL server to stay failed over until the SQL service o the backup is killed. An odd requirement but that's what I got.

My thought is to use a second probe against some sort of debar parameter which the DBAs will create.

Thank you

Cathy

Syed Iftekhar Ahmed · ‎08-20-2008

I have never come across a config that will keep backup serverfarm remain active after active recovers.

I am sure you know that Backup serverfarms keeps serving the existing connections when the Active serverfarm recovers.

If you use stickiness with backup serverfarm the active clients will remain connected to the backup serverfarm (until the sticky timeout expires)

The new connections (if sticky is not configured / if sticky is configured and client is not in sticky db) will always go to the recovered "active server farm".

The change we got in (3.0(0)A2(1.1A) ws the introduction of "sticky" key word for backup serverfarm. Now instead of Backup serverfarm automatically inheriting the stickiness from primary serverfarm, we have an option to disable stickiness for backup serverfarm .Thus making it possible for new connections to go back to Primary server farm when it comes back (instead of waiting for sticky timeout to expire on backup Server farm).

Syed

View solution in original post

Syed Iftekhar Ahmed · ‎08-20-2008

I have never come across a config that will keep backup serverfarm remain active after active recovers.

I am sure you know that Backup serverfarms keeps serving the existing connections when the Active serverfarm recovers.

If you use stickiness with backup serverfarm the active clients will remain connected to the backup serverfarm (until the sticky timeout expires)

The new connections (if sticky is not configured / if sticky is configured and client is not in sticky db) will always go to the recovered "active server farm".

The change we got in (3.0(0)A2(1.1A) ws the introduction of "sticky" key word for backup serverfarm. Now instead of Backup serverfarm automatically inheriting the stickiness from primary serverfarm, we have an option to disable stickiness for backup serverfarm .Thus making it possible for new connections to go back to Primary server farm when it comes back (instead of waiting for sticky timeout to expire on backup Server farm).

Syed

ciscocsoc · ‎08-20-2008

Thank you Syed,

I'll just have to get creative with a secondary probe :-)

Thank you for confirming that I hadn't missed an obvious method.

Kind Regards

Cathy