[Info-vax] ANOTHER minicopy question

[Keith Parris]

On 1/3/2013 2:32 PM, Phillip Helbig---undress to reply wrote:
 > In article <op.wqcqkxo8i0cpeq at kparris4.americas.hpqcorp.net>, "Keith
 > Parris" <keith.parris at hp.com> writes:
 >> 2) Boot each node by default as a satellite node of the other 2 nodes.
 >> This allows the booting node to boot from the present 2-member 
system disk
 >> shadowset instead of its out-of-date local copy, and then once it's 
up you
 >> could add its member to the system disk shadowset with a Mini-Copy
 >> operation.
 >
 > I have never thought of that.

I hadn't either, until it was pointed out to me. I told customers for 
years that you couldn't have a system disk shadowed between sites if the 
only access to a member disk was through the VMS MSCP server on a node 
at that site. Then a customer told me not only was it possible, but they 
had been doing it for years in their 3-site DT cluster, with a 3-member 
system disk shadowset, one member at each site, MSCP-served to the 
cluster, with nodes at each site set up by default to reboot as 
satellite nodes from the cluster as a whole.

With disks on SAN and SAN Extension between sites, system disks shadowed 
between sites became trivial to accomplish, since you didn't have to go 
through the VMS MSCP server for access to the member disks at each site, 
and any surviving VMS node at any site could keep all the shadowset 
members up-to-date.

 > However, I should mention that both
 > members have direct connections to only one node, the node which boots
 > from them.

You mention "both menmbers" but I was thinking, based on earlier 
contents of this thread, that there would be a 3-node cluster with 
3-member system disk shadowset, with each member being a local disk 
accessible only through the VMS MSCP Server on its local node.

The combination of AMCVP and booting by default as satellite nodes from 
the remaining nodes in the cluster should make this configuration workable.

Where it would get tricky is if you lost 2 nodes out of 3. Then you'd 
have to use something like Availability Manager or a console IPC>Q 
command to do a quorum adjustment on the sole remaining node; then the 
other 2 nodes should be able to reboot from it. You would need at least 
two Multi-use bitmaps on each node for AMCVP, one for each of the two 
Mini-Copies which would be needed to recover the two different system 
disk members (which would leave the shadowset at different times).