home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!gatech!cae!cae!not-for-mail
- From: chris@cad.gatech.edu (Chris McClellen)
- Newsgroups: comp.os.os2.misc
- Subject: Re: Did microsoft really write an OS/2 "Terminator" crash prog?
- Date: 30 Dec 1992 16:07:13 -0500
- Organization: Georgia Institute of Technology, CAE/CAD Lab
- Lines: 54
- Message-ID: <1ht321INNdfe@cae.cad.gatech.edu>
- References: <1992Dec29.052115.27441@cbnews.cb.att.com> <1992Dec30.172609.17583@spartan.ac.BrockU.CA> <1hspsiINNc1d@cae.cad.gatech.edu> <8252@lib.tmc.edu>
- NNTP-Posting-Host: cae.cad.gatech.edu
-
- In <8252@lib.tmc.edu> jmaynard@oac.hsc.uth.tmc.edu (Jay Maynard) writes:
-
- >In article <1hspsiINNc1d@cae.cad.gatech.edu> chris@cad.gatech.edu (Chris McClellen) writes:
- [ low tolerance example deleted ]
-
- ]I assume you're referring to MVS. The action you describe isn't as
- ]unreasonable as it sounds, for at least two reasons:
- ]1) Messages appearing on the console are assumed to be important enough that
- ]they should be seen by a human. Thus, they are buffered until they're
- ]displayed. There are a finite number of those buffers, and there's not a lot
- ]that can be reasonably done when those are exhausted.
-
- This is true. However, things like disconnecting a terminak
- (non console) can disable that terminal onsome of their systems
- until next reboot. Booting IBM's larger systems can take a long time.
- Secondly, if the messages are not critical, a lost console shouldn't
- really bring the system down, especially if its just some
- messages about print jobs being completed, or even aborted. Disk
- errors sure, but not normal status messages, etc. But alas,
- its not really an issue here, that is just a bit of a personal
- opinion of mine... Story: one engineer was working somewhere
- round here on a mainframe (I believe it was one of the IBM boxes).
- The console cord was draped carelessly across the floor.. he dragged
- his feet, so when he waltzed by, yes, he tore the cord out, and
- the machine shut down. Big pain -- and I dont believe it was
- necessary for the machine to have shut down. again, a non issue..
- Just stating some of IBM's fault tolerance...
-
-
- ]2) There is a facility for disabling a console that goes haywire if another
- ]console is available. If there are no consoles available, the system can't be
- ]operated anyway, since there is no means for the operator to interact with the
- ]OS.
-
- Also, while this is true, I was only giving examples of how their
- thoughts seem to have influenced OS/2. Frankly, I was surprised
- that when I turned off my monitor, OS/2 didnt crash (:) -- humor).
-
- Seriously, when talking about a SINGLE USER OS, on a microcomputer,
- you really dont need to design with a mainframe attitude.
-
- The poster I replied to asked why OS/2 just doesn't have a set
- of defaults, so if config.sys is not found, it boots anyway.
- I dont want to discuss why or why not it may be good programming
- practice, but I was just saying that OS/2 seems to have gotten
- some qualities from various IBM mainframe/minicomp OS's.
-
- One one machine with an 8 inch floppy drive, I forgot to put in a
- floppy before I ltry to access the drive. I had to reboot
- the machine, because the OS disabled the floppy device
- since it detected an error on read.
-
-
-
-