[Info-vax] Cluster node fails to boot from shadow-set member

[Volker Halle]

Richard,

the reason for the SHADBOOTFAIL Crash seems to be hidden/overwritten by the CONFIGSCAN message on the console.

%SHADOW-F-...

Starting with OpenVMS Alpha V6.2, there is an additional message printed on the
console terminal together with the bugcheck printout, which contains additional
information about the reason for this bugcheck.

Node 2 seems to boot from the physical disk $2$DKA0: and forms a DSA0: System disk shadowset with a single shadowset member: $2$DKA0:

Node 1 then boots from it's local physical disk $1$DKA0: and tries to also form a DSA0: shadowed System disk DSA0: with a member of $1$DKA0: - after joining the Cluster, it detects, that there already is a DSA0: shadowset with a single VALID member of $2$DKA0:. 

Node 1 then crashes with SHADBOOTFAIL and the most likely reason:

%SHADOW-F-BDNOMBREX, boot device is not a member of existing shadow set

If you want the 2 rx2600 i2 Systems to boot from the same shadowed System disk DSA0: you need to use a shared SCSI bus via which both Systems can directly access $1DKA0: and $2$DKA0:

Volker.