Showing results for 
Search instead for 
Did you mean: 



I want to have some best practices for implementing SFHA 6.0 on a 2-node T5-2 (Control Domain, Guest LDOM).

Basically we would want to have a 2-node cluster on the Guest LDOMs (not on the control domains) and  we would install VCS in the Guest LDOMs.

For the Storage Foundation portion (VxVM/VxFS), we would like to install the whole SF stack on the Guest LDOMs, the control domainwill be just exporting the disk devices to the Guest LDOMs (note there is no SF installed in the Control Domain).

What we found out, exporting the physical disks (which are on different paths e.g c4 & c6) from the control domain to the guest domain, result in the virtual disks in the Guest Domain being under the same controller (e.g c1). That is all the disks fall under the c1 controller in the Guest Domain, even though they are exported using different physical paths in the backend and coming from different HBAs.

So my questions are:

- Does the path failover still function despite that the Guest LDOM can only see 1 path to the LUN (and not 2 paths, as presented to the Control Domain)

- Is there a way for the Guest LDOM to see the LUNs on 2 different control numbers?

Remember we do not have SF installed in the Control Domains.


Thank you in anticipation


1 Solution

Accepted Solutions
Accepted Solution!


3 Replies

hi Danilo, Have you refered

hi Danilo,

Have you refered to

It is possible to export the individual paths to guest domain for each disk and DMP do path management. Please refer to section "Guest-based Storage Foundation stack model" in the above document.





Accepted Solution!


Cool stuff! Many thanks

Cool stuff!

Many thanks Venkatareddy!