[Info-vax] Cluster node fails to boot from shadow-set member
Stephen Hoffman
seaohveh at hoffmanlabs.invalid
Wed Apr 17 12:19:07 EDT 2019
On 2019-04-17 14:29:18 +0000, Richard B said:
> So, in the absence of a shared SCSI bus, which I do not have, what are
> the possible solutions? Non-shadowed system disks?
For a cluster member to boot... You need a direct and multi-host path
to the system device storage and here that'd usually be either be a
multi-host supported and Tagged Command Queuing (TCQ) -capable SCSI
controller and supported shared SCSI storage such as an MSA30-MI HDD or
SSD shelf or would be a Fibre Channel Host Bus Adapter (HBA) and Fibre
Channel storage, or you could configure this Integrity box as a
satellite and a served satellite path and network boot and the OpenVMS
host serving the disks can shadow or RAID, or you can configure
multiple parallel system disks. Here, you're likely going to have
parallel system disks and with these then mounting a storage device and
that usually in SYLOGICALS.COM storage for a common set of shared
files—see SYLOGICALS.TEMPLATE for a partial list of some of the roughly
two-dozen files necessarily shared for a stable and consistent cluster.
The shared storage for SYSUAF, RIGHTSLIST and the rest of the
necessarily-shared files can be configured with host-based or
controller-based RAID. The local system boot device storage will
probably be controller-based RAID.
--
Pure Personal Opinion | HoffmanLabs LLC
More information about the Info-vax
mailing list