Hello,
Are you interested in improving service availability? If so, you can help...
Since 1995 I have had a need for SAFE concurrent read/write access by
multiple systems to HSZ40 devices. This would solve our performance,
reliability, and availability problems with using NFS. This capability
means no single system would control a disk as does NFS and TruCluster
(aka, DECsafe, DEC ASE). Hence availability to the disk would be
continuous to all other systems when a single system fails. Currently, NFS
clients wait while the NFS server is out of service and TruCluster
configurations suffer a failover period -- TruCluster introduces
additional headaches.
We have a 4100 and 2100A accessing SW300/SW800 with dual-redundant HSZ40s
via KZPSAs in a DECsafe hardware configuration.
I posted the question twice in the past and the word from Digital is
the capability is coming. What they meant at the time was DECsafe.
Now I am working with Digital Consulting to have the capability added
to Digital UNIX but Digital does not appear to think customers would be
interested in this capability. ARE YOU? If so, please let me know to
help build a case.
Some folks at Digital think it is fine to have a disk unavailable during
a failover period but I argue it is better to have the disk available
all the time.
--
Regards,
Richard Jackson
Computer Center Lead Engineer
Mgr, Central Systems & Dept. UNIX Consulting
University Computing & Information Systems (UCIS)
George Mason University, Fairfax, Virginia
Received on Fri Oct 03 1997 - 15:33:13 NZST