LVM mirror performance impact with fail disk

[OSG]

Hi
I come here for some idea about some dysfunctions on Hight Availabiliy HPUX configuration.
I have a cluster MC/SG with 2 HPUX servers L3000 in hpux11.0.
We have 2 Virtual Array (VA7100) in a redondant SAN loop configuration.
An oracle database is in a mirrored VG by LVM between these 2 VA bays, with PV Link and MC/SG between the 2 HPUX servers.

After a human error one VA bay was improperly reboot and device become inaccessible, in a strange status.
The Mirror work and we could see the vg in a state
Available/stale
But after we had very bad performance in I/O transaction for the database.
So we had rebooted the entire cluster (servers and bays) for recovery better performance!!

I think than the LVM mirror was the cause of this! The LVM try to write on the fail disk and wait for time out for each I/O.
And the best way to solve this is to reduce the mirror and after to repare the fail Virtual array !

Anyone have the same problèm with HPUX and mirror??
Thank for your help to understand how improve the availabity for my cluster !!

[Torsten.]

see
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1137320