On Nov 9, 2005, at 5:17 PM, Whoever Whatever wrote:
Hi,
I wonder if anyone can answer this question, we are using sun Storedge
3510FC on about ten different systems, one of the setup(all but
one) used
three of those unit linked with 2 Gb fiber loops to see maxium of 3
x (12 x
73GB) hard drives, the controllers are dual channels to the arrays.
After
configuration, they all created /dev/dsk/ and /dev/rdsk/ with WWN, ie
/dev/rdsk/c7t40d0s20440FBDE0Fs2. We configured one with just a
single array,
that's 12 drives wtih 73GB each, all devices show up in /dev/dsk and
/dev/rdsk without WWN, ie /dev/rdsk/c7t40d0s2, c7t40d1s2 and etc.
Both setup
worked fine, they used the same hardware, same version of Solaris,
even the
same release date, applied the same patched, but don't know why the
system
with single arrary has normal device node and the others has the
weird World
Wide Name. I search the internet, found only one post talk about
conflict
Hard Drive ID on the SAN will cause that, but it will funtion as
normal when
addressed with the WWN node, I tried reconfigure with "boot -r" and
devfsadm
on the system with single arrary but not able to cause it to create
WWN
node, anyone know what's the real cause of this?
It sounds like that the box where you're seeing the controller and
disk doesn't have multipathing turned on.
In /kernel/drv/scsi_vhci.conf the default setting for mpxio-disable
is “yes”, this must be changed to “no”;
mpxio-disable=”no”;
and the default the setting for auto-fail back is “disable”, you want
to change this to “enable”;
auto-fail back=“enable”
You must also reboot for these changes to take effect (this is one of
the few times you have to do that in Solaris).
After the reboot, you should see the WWN.
Joe
