Hi,
for some application running on Solaris we needed to purchase Sun StorEdge
3310. It has 2 SCSI controllers and 5 disks (2disks x2 for mirror, and one
for hot spare).
Now we'd like to test redundancy of this storage config.
Are we allowed to physically detach disk so we can observer 5th disk used ?
I'd like to be sure that in case of disaster failover will work fine (either
with failed disk, or failed controller).
How could I test it so we can be sure ?
We tested failovering to standby scsi controller, with following commands:
sccli> show redundancy

sccli> fail primary

failover to standby controller takes more that a minute and during this time
writing to directory where storage disk is mounted is not possible. Is this
normal or should it take much shorter time ?

thanks a lot,
jmiklo

In comp.unix.solaris jmiklo <jmiklo@hotmail.com> wrote:
> for some application running on Solaris we needed to purchase Sun StorEdge
> 3310. It has 2 SCSI controllers and 5 disks (2disks x2 for mirror, and one
> for hot spare).
> Now we'd like to test redundancy of this storage config.
> Are we allowed to physically detach disk so we can observer 5th disk used ?

It's not really recommended, but it won't break anything.

> sccli> fail primary
>
> failover to standby controller takes more that a minute and during this time

No. It should happen within less than a second. Not sure what would cause
it to take so long for the failover.

Scott