Yes, when the remote horcm daemon is not accessible, an SVOL-SSUS takeover must be done. This is normal Truecopy behavior in the case of a primary site failure or fibre link failure and is documented in the Truecopy documentation, as is the recovery procedure of running pairresync -swap.
We've received several questions lately about this exact same failure scenario, and while I'm reluctant to reproduce information that is described in agonizing detail in the Truecopy documentation, I'm starting to think we may have to duplicate this information in the Sun Cluster docs. What do you think? stephen On 07/ 5/09 07:49 PM, Sergei Kolodka wrote: > Guys, > > I've got question about HORCM behaviour on SC, or any cluster... > > If P-VOL host is down and S-VOL HORCM instance is running should horctakeover > be able to actually take control over array or all it can/should do is to put > S-VOL into SSUS Writable state and let admin to do pairresync manually after > that ? > > The problem we encountered is - if we do fallback to first node without doing > manually pairresync while we're on second node first cluster screws itself up > and can not start anything before we split pair. Is this behaviour expected > and is it should be exactly that way ? > > We're running SC3.2 non-geo with TrueCopy replication. > > Thanks in advance. >
