Hi Managers,
We are about to install new alphaservers (Tru64 Unix 4.0-F) with some
storage attached to it. For the storage we are planning a dual HSZ70
configuration in transparent failover mode. The disks will be configured in
a Raid 5 configuration. Filesystem will be advanced file system.
We are looking into a way to make this configuration as robust as possible.
Because Tru64 4.0-F does not support a multibus failover configuration we
had to choose transparent failover mode for the HSZ controllers. We are
planning to build four Raid 5 sets. Two of them (lets say unit d100 en
d300) will have HSZ1 as their primary controller and HSZ2 as their failover
controller. The other two (lets say unit d200 and d400) will have HSZ2 as
their primary controller and HSZ1 as their failover controller. In case of a
failure of one of the HSZ controllers all units will be serviced by the
other controller. Our main challenge is to make disks as quickly (and as
simple as possible) reachable through the other controllers. The main
problem of course is the change in device file naming when a unit fails over
from one to another HSZ controller. Does any of you has any experience or
ideas on how to accomplish this in the best way. We were thinking to write a
little script to change links in the /etc/fdmns/<domain> directories, but I
am wondering if this can be done in any other way.
Regards,
Gert-Jan Hilbrands
Received on Fri Jun 15 2001 - 11:51:10 NZST