[Info-vax] Exploring a hard drive on a new Vaxstation

[Stephen Hoffman]

On 2018-01-29 19:09:00 +0000, Rod Regier said:

> The 21 bit addressing limit could give you "interesting" issues.
> 
> As long as the blocks needed for the *boot* were low addresses on the 
> drive, you could use a very large drive.
> 
> But heaven help you if the needed blocks ended up on high addresses.  
> In that case, the game was over unless you performed a dump/reload for 
> the data to push those blocks back down again.


Boot failures are rather more benign than what can happen here with 
VAXstation 3100, MicroVAX 3100 models 10 and 20, and MicroVAX 3100 
models 10e and 20e below console VMB V6.4, when configured with a boot 
disk larger than ~1.073 GB.

The system dump will corrupt the system disk on crash, just as soon as 
one or more of the extents of a dump-related file happen to be located 
or relocated out beyond 1fffff (base16) blocks.  The write I/O will 
wrap around.

That wrap will stomp on some matching lower-addressed range of disk blocks.

That's a corruption which can make any subsequent system operations 
rather less than entirely reliable, too.

Data disks have different and larger addressing limits, and those 
limits based on the particular VMS or OpenVMS VAX version in use.




-- 
Pure Personal Opinion | HoffmanLabs LLC