| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 558.1 | | SUTRA::16.36.2.83::Bats | Speeding, speeding, I'm always speeding | Tue Jan 14 1997 13:12 | 6 |
| 558.2 | SCSI cable ? | CIVPR1::SIMMONS | Mike Simmons (301) 918-5597 | Mon Jan 27 1997 23:13 | 5 |
| > So use a RAID410 or RAID450.
> That has failover capability of the RAID controller itself.
I don't mean to be picky, but isn't the SCSI cable(s) still a single point of
failure?
|
| 558.3 | Multi-Bus Failover Mode | DECWET::LEES | Will, NTSG DECwest, Seattle | Tue Jan 28 1997 15:47 | 12 |
| Yes, if termination is broken anywhere on the Scsi cable potentially both
systems could hang.
One solution to this is a special cabling configuration of the RAID subsystem
called "Multi-bus failover mode". In this configuration there are two disjoint
Scsi buses, each between a host and the RAID subsystem. Each one is
electrically separate; a break in one does not effect the other.
Support for MBFM is dependent on a RAID firmware change that is not expected
until Summer.
Will
|
| 558.4 | MBFB: not under Windows NT? | CIVPR1::SIMMONS | Mike Simmons (301) 918-5597 | Wed Jan 29 1997 19:51 | 11 |
| > One solution to this is a special cabling configuration of the RAID subsystem
> called "Multi-bus failover mode". In this configuration there are two disjoi*
> Scsi buses, each between a host and the RAID subsystem. Each one is
> electrically separate; a break in one does not effect the other.
> Support for MBFM is dependent on a RAID firmware change that is not expected
> until Summer.
I didn't think there was (currently-or in the near future) any support for
this under Windows NT (with, or without Clusters). Am I mistaken about this?
|