Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!hellgate.utah.edu!helios.ee.lbl.gov!beva.bev.lbl.gov!wbrown
From: wbrown@beva.bev.lbl.gov (Bill Brown)
Newsgroups: comp.realtime
Subject: Re: rebooting hung vxWorks system
Message-ID: <4392@helios.ee.lbl.gov>
Date: 6 Dec 89 15:21:52 GMT
References: <5321@ncar.ucar.edu> <41645@srcsip.UUCP>
Sender: usenet@helios.ee.lbl.gov
Reply-To: wbrown@beva.bev.lbl.gov (Bill Brown)
Distribution: na
Organization: Lawrence Berkeley Laboratory, Berkeley
Lines: 33
X-Local-Date: 6 Dec 89 07:21:52 PST

In article <41645@srcsip.UUCP> sbrandt@src.honeywell.com (Scott Brandt) writes:
>In article <5321@ncar.ucar.edu> martin@aster.UCAR.EDU (Charlie Martin) writes:
>>Has anybody devised a method for remotely booting 
>>(ie. over the network) a vxWorks system which has crashed?
>
>In answer to your question:
>	If the system has crashed, it would probably be impossible to talk to 
>it to tell it to reboot.  Therefore you would need some external way to toggle 
>the reset line in the chassis.  One of my hardware-hacking coworkers 
>connected a device to one of our vxWorks systems that toggles the reset line
>if the phone rings more than 8 times in the lab.  This is hardly an elegant
>solution, but in order to remotely reboot the system you will need something
>analogous to this.  If you are creative you can probably put some sort of 
>device on whichever host your vxWorks system boots off of to accomplish 
>the same thing.  

We've developed a dead-man system that seems to work fairly well.  It's our
intent to release it to the archive when we're sure it's working.  It
uses the built-in deadman timer on the MVME-147 and also is being moved
to the '133 and '135.  The scheme depends on the task-scheduling invoking
a function that keeps the dead-man from timing out.  The problem is that
all 3 boards have different hardware, so most of the code is system-dependent.
Probably we'll add stuff to the vxWorks system-dependent code.

I certainly wouldn't want it to stand between me and the end of the world,
but it does seem to take care of quite a few system hangs.

							-bill
							wlbrown@lbl.gov

Disclaimer:  These opinions are my own and have nothing to do with the
    official policy or management of L.B.L, who probably couldn't care 
    less about employees who play with trains.