[Info-vax] Why it is a good idea that OpenVMS isn't on x86-64 just yet

[terry-groups at glaver.org]

On Friday, January 15, 2016 at 9:42:36 AM UTC-5, johnwa... at yahoo.co.uk wrote:
> On the other hand, over the decades, VMS systems have had a variety of
> obscure hardwareish bugs, some of which have reached real customers
> in the field. VMS has a variety of system level tools which can be 
> used to record (and later analyse) portions of system state when
> things misbehave. There used to be (and maybe still are) people who
> could look at crash dumps and such like and shed some light on what
> had been happening.

Probably the most complete hardware snapshotting (I don't know about the 9000-series, from the same development group) was the VAX 8600 (and later the 8650). The "console media" was expanded from the floppy / TU58 of earlier VAXen to an RL02. In addition to the normal things on the console media, there was room for 2 (IIRC) hardware snapshots. My memory (which, unfortunately, lacks ECC) tells me that each MCA (the logic chips with the "double decker hat" heatsinks) had an accompanying SIP module which collected the relevant logic states and shipped them serially to the console processor.

In addition to finding design issues, this eventually became useful for locating hardware faults - once enough data about "if X fault happens, replacing the Y board fixes it Z percent of the time" was collected, DEC could interpret the snapshot, rather than just replacing boards one-at-a-time.

I don't know if the logic snapshot facility was part of the original design or a later add-on. Certainly the 8600 had a troubled enough gestation that it was a useful feature for the designers. I believe the 8650 provided the originally-planned 8600 performance. If the rest of the system was up-to-rev, it was a simple 2-board swap to upgrade 8600 -> 8650.