SUMMARY: SC 3.1 + SDS metaset-problem: drive is not common with other host

From: <>
Date: Wed Jan 03 2007 - 05:19:37 EST
<Original posting below>

Hi agn,

and thanks to Graeme Elsworthy [] for his response, and
esp. to prabir sarkar [] who pointed me in the right
Prabir suggested to remove and re-attach the disk in the following way:

- with luxadm/cfgadm remove the disk from both the nodes one after other.
- run devfsadm -C, devfsadm on both nodes, scdidadm -r on either node.
- with cfgadm configure the disk back on both the nodes.
- run devfsadm -C, devfsadm on both nodes, scdidadm -r on either node.

I removed the disk with cfgadm -c unconfigure, devfsadm + scdidadm, and voila,
it went away! Unfortunately, I was unable to add it back with cfgadm, so,
Prabir suggested a reconfigure boot which did the trick.
The disk came back as did/d9 on both nodes, and I was able to add it to
diskset and build the metadevices etc.

Thanks again,

and have a nice hackin',


Hi gurus,

I've got a Sun cluster consisting of two X4200 and two SE3310, running
Sol10 x86 and SC3.1 U4, patch rev. is EIS-DVD 12/06.
I have defined two LUNs on each SE 3310, and I'd like to map them to two
disksets, one called "log-dg" and the other one called "ora-dg". The
problem is, that I can't add the second "disk" to the diskset "ora-dg",
since SVM complains that this drive is not common with the other host.
But, I'm quite sure it is, format etc. shows the drive, and even
scdidadm -L shows it without any problems.
One further information: When I first booted the machines, the storages
had a different configuration, and I re-configured them. After this, the
disk in question showed up as d9 on node 1, and as d11 on node 2 - not
very nice for a shared device, so, since I was unable to get rid of this
mappings (even a reconfigure boot after completely disconnecting the
storages helped), I booted up the nodes in non-cluster mode and
edited /dev/cluster/ccr/did_instances by hand.
That helped, after this and a devfsadm && scdidadm -r, the drive shows
up as d9 on both nodes, as I expected. But, see above, the cluster still
thinks these are different drives (although did_instances seems to be

Any clues?

Thanks in advance,

sunmanagers mailing list <>
sunmanagers mailing list
Received on Wed Jan 3 05:20:16 2007

This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:44:03 EST