Article 144531 of comp.os.vms: I was reading the upgrade manual in preparation for doing a VMS upgrade and I got to the part about doing a rolling cluster upgrade. When it talked about maintaining quorum in the cluster, it said: : the cluster. If the cluster hangs during a shutdown, : enter the following commands on the system console : of a CPU that is still a cluster member: : : $ : >>> H : >>> D/I 14 C : >>> C : IPC> Q : IPC> As for the specifics of what, how, and why this is, the manual is silent. Can anyone explain what this is doing? [I presume it is to adjust the quorum on a cluster that has already lost quorum and hung.] : $ Put processor into console mode. : >>> H Halt CPU. (On all recent VAXes, this is a no-op - console mode is implemented by VAX code in ROM, so the VAX is either running the console code, or running VMS. But some older machines - going back to the original VAX 11/780 - had a separate front-end processor - an LSI-11, in the case of the 780.) : >>> D/I 14 C Deposit the hex value C into internal processor register (IPR) 14. This IPR is the SIRR, the Software Interrupt Request Register. Depositing a value in this register request a software interrupt at the indicated priority level. : >>> C Continue. The requested interrupt - at software priority hex C, or 12 - takes place as soon as there are no higher-priority interrupts pending. : IPC> Q The interrupt handler for priority 12 is the IPC routine, which implements a (very) small set of commands. Normal timesharing is now blocked. Q requests quorum re-calculation of quorum based on the nodes currently accessible. Votes from nodes that had previously been in the cluster, but have since left, are "forgotten". This is the only way to *lower* the quorum. : IPC> Exit the IPC routine, and continue with normal timesharing. -- Jerry