We have one HPUX N4000 server connected to two brocade silkworm 2800 switches both going to one VA7400 SAN with two controllers all fibre connected.

We have been doing fail over tests with this set up to see if we loose one controller or a brocade switch connected to the VA7400 it, should, go over to the alternate path.

If we do a vgdisplay -v /dev/vgx the screen does show an alternate path, and the correct one, but the system still does not fail over. Our database crashes.

We always crash because we lose contact with VG06 holding our logicaldbs. Rootdbs and physicaldbs are on another volume group

We only run Informix Database ver 7.3 on this system.
We have altered the pvtime out settings going from 30 to 180 seconds and many inbetween but with no success.
Is there a setting, possibly, in Informix to let it know if it cannot get to the database to use an alternate path .
I thought this should be solved with the Operating system but apparently not.

We run async