🚀 go-pugleaf

On Wed, 24 Jun 1992 01:30:00 EDT Lee Garner said:
>On Tue, 23 Jun 1992 20:57:34 PDT Ed Rovera said:
>>   One of our clients noted that the WAIT option on the VMTAPE MOUNT
>>command doesn't allow for a timeout value.  ...
>
>  Actually, there is a timeout value.  It just isn't implemented where
>your client expects it to be.  Take a look at the "REMINDER" statement
>in the VMTAPE CONFIG file.
>
>  We have "REMINDER 30 900 60" in our config file, which gives us a
>maximum waiting period of 15 minutes before VMTAPE cancels the mount.
>And, "Yes", our operators hate it and sometimes mounts get canceled
>on the evening and weekend shifts.  :-)
>
>> ...
>>   Does anybody else think that the ability to specify a timeout on
>>the WAIT option would be a valuable addition to VMTAPE's MOUNT command
>>or is UCSF the only place where one has problems getting a tape
>>mounted on weekends?
>
>  Sure, it would be useful.  The site could set a global default
>waiting period in the VMTAPE CONFIG file as above, and then if the
>user didn't want to wait that long (or didn't mind waiting even
>longer) then s/he could override it with the command option.  Sounds
>neat to me!

   I think that the REMINDER statement is fine for setting a *maximum*
time that a site will allow a mount to remain pending before it wants
the request considered 'unanswered' or 'unanswerable.' I'm a bit leary
of having users override this and have a mount pending for an entire
day but that's something that we can debate at another time.  Our
problem with using the value in the CONFIG statement is that the user
can't set a *lower* value for his expressed reason.  Our REMINDER
statement reads 'REMINDER 60 3600 60' so we have stated that we want
VMTAPE to wait a maximum of 60 minutes NO MATTER WHAT.  If the user
wants to have his machine wait only 45 minutes -- for whatever reason
-- he doesn't have the option.  I'm in complete agreement that the
REMINDER statement constitutes the default timeout but my client is of
the opinion that making his machine wait 60 minutes when he'd rather
it enter an error routine sooner is a bit Draconian and I see his
point.

                                      - Ed Rovera
                                         UCSF ITS Division
                                         415-476-3119