we've followed the guidelines of the Oralce White paper "Integrating Oracle VM into a Disaster Recovery Solution using SAN".
So far so good except that we can't do it twice. I.E. we can't do repetitive tests of that recovery procedure.
We can't gain access of a repository (or a clone of it) that had been already mounted once
This is defeating the aforementioned white paper but Oracle is not very reactive to acknowledge that problem/bug/feature
I have an open SR on that which has been waiting for Oracle internal news for a month
Reproducible failure :
- Create a new repository (and do nothing more in there) - Release Ownership of that repository - close the iSCSI connection to that LUN - re-connect that iSCI LUN - refresh the server ---> Failure : OVMAPI_6055E Discover target: xxxxxxx, Discover failed with error: java.lang.NullPointerException at com.oracle.ovm.mgr.api.util.StoragePluginUtils.detectFileSystemSizeMismatch(StoragePluginUtils.java:132)
If it matters, we are using "unamanaged iSCSI storage" (Equallogic)
We've found a workaround : Creating a new server pool allows to gain access to the repository.