[HECnet] Tops20 uptime..

[Thomas DeBellis]

That is not too shabby at all!  So let's see, that means you booted on 
Thursday, September 9, 2021 at 8:13:49 am, local. Right?  And that works 
out to be 19 weeks, two hours, 7 minutes and 53 seconds (assuming I 
haven't broken uptime tinkering with this).  VENTI2:: (my other 
development 20) has been up for 5106:06:49 or 30 Weeks, 2 Days, 18 
Hours, 6 Minutes, 45 Seconds and 675 Milliseconds, which isn't bad as 
this is what I run all my fiendish experiments on.

The exclamations after the uptime quantity suggest that you are running 
what appears to be a Stanford based EXEC as I do not recall that they 
exist in the standard DEC release.  I have a few questions for you, if 
you are willing:

 1. Have any Systems Concepts modifications handled the previously
    mentioned 35 bit uptime limit?  In other words, can you still get an
    UP2LNG BUGHLT's?
 2. Are you using queued shutdowns (I.E., the PANDA ACJ)?
 3. What is the maximum virtual address space?  The maximum for a KL is
    32 sections, so I had been wondering.

The problem I have with the PANDA ACJ is that it determines what to 
queue and what not to.  That is not unreasonable given that the code 
works by hooking HSYS% and there is no interface for the JSYS to specify 
whether to queue or not.  The problem is that if you set a downtime to 
shutdown before an UP2LNG BUGHLT and subsequently crash before then, the 
ACJ will set the same downtime when you reboot, which is probably not 
the right thing to do.

So from TOMMYT::, one sees:

    @iNFORMATION downTIME
    Shutdown Time:            Up Again: Reason:
    Fri 18-Feb-2022 12:30:00PM-EST    12:45:00PM-EST         UP2LNG BUGHLT
    @

and from Galaxy:

    OPR>shoW staTUS
    OPR>
    13:53:44        -- System Device Status --

    [...]

    ** Timesharing will cease in 694 hours and 36 minutes*

The unfortunate limitation here is that while Tops-20 Galaxy will 
display a shutdown time, there is no code to change it.  You have to 
exit and use the EXEC ^Ecease command.  There was some command that an 
FE with maintenance capability could issue, but it couldn't have been 
through the EXEC, which only allows an enabled WHEEL or OPERATOR to do 
that, despite the monitor allowing maintenance.  Since I couldn't 
remember that particular program, I changed the PANDA EXEC. Anybody 
remember what it might have been?

I think the Tops-10 approach of using Galaxy to handle system shutdowns 
is probably superior or could be made to be.

So, fingers still crossed for the 18th...

> ------------------------------------------------------------------------
> On 1/20/22 10:23 AM, Peter Lothberg wrote:
>
> SC40 SOL::
>
> $sys
>   Thu 20-Jan-2022 10:21:42  Up 3914:07:53!!!
>   1+8 Jobs   Load av   0.01   0.02   0.02
>
> Nb, this is real HW...
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.sonic.net/mailman/private/hecnet-list/attachments/20220120/d96053aa/attachment-0001.htm>