[Info-vax] 8.4 and DECwindows CDE login box not coming up

[johnwallace4 at yahoo.co.uk]

On Saturday, 23 May 2015 15:35:41 UTC+1, Stephen Hoffman  wrote:
> On 2015-05-23 13:58:37 +0000, johnwallace4 at yahoo.co.uk said:
> 
> > Picking semi-random system parameters and/or account quotas and 
> > changing them semi-randomly might work. Or it might not.
> 
> Wasting sixteen pages of memory by picking an unnecessarily large value 
> such as 4,194,304 is not usually a big deal.
> Or if sixteen pages is a problem, then -- in the ancient mantra of DECUS 
> symposia long past -- "buy more memory".
> 
> 
> -- 
> Pure Personal Opinion | HoffmanLabs LLC

That's not really semi-random though, it's got very sensible logic
and numbers and arithmetic and VMS internals knowledge behind it, and
is a very sensible suggestion, where applicable. 

The parts of the picture which we have from Phillip suggest that he
thinks he's got plenty of memory, so hopefully the suggestion is
applicable here. Can hardly hurt to try it; the days when things
like GBLPAGES and GBLSECTIONS were much of a constraint should have
gone away a long long time ago.

It's not Phillips opinion that counts, though, it's not our opinion,
it's VMS's opinion. 

Am I misremembering or do we have such a thing as a "SHOW SYSTEM"
output on the failing node, taken at the time when DECwindows
startup is hanging?

That wouldn't necessarily even need to be done from the failing
system, courtesy of SHOW SYSTEM /node=whoever. If it works, we 
get some info we don't currently have? If it doesn't work, we
may get a hint as to what the broader issue might be.

Or is that a daft suggestion?

Here's a quite daft (but hypothetically possible) suggestion:
if a crash dump was forced at the point when DECwindows is just
sitting there, would anybody here be volunteering to look at the
dump, for entertainment? (Not me, sadly I don't have the skills
for that task).

Meanwhile, we wait patiently...