[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: watchdog
- To: tass@wwa.com
- Subject: Re: watchdog
- From: Chris Albertson <chris@topdog.pas1.logicon.com>
- Date: Tue, 19 Aug 1997 13:14:15 -0700
- Old-Return-Path: <chris@topdog.logicon.com>
- Organization: Logicon RDA
- References: <9708182100.AA14362@runner.nofs.navy.mil>
- Resent-Date: Tue, 19 Aug 1997 16:15:48 -0400 (EDT)
- Resent-From: tass@wwa.com
- Resent-Message-ID: <"dsjWzC.A.2o.u6f-z"@kani.wwa.com>
- Resent-Sender: tass-request@wwa.com
- Sender: chris@topdog.pas1.logicon.com
aah@nofs.navy.mil wrote:
> then restart properly. My preference would be to protect with hardware
> against any damage to the camera or electronics, and then just suffer
> the infrequent 'crash'. Adding an UPS is a small pertubation that does
> guard against the seconds-duration power outage that often occurs during
> a storm, but I'd rather have to restart a system when the power dies for
> an hour or two (and the UPS gives up) than have to design in a smart
> recovery system. The loss of a night's worth of data is acceptible for
> an automatic telescope.
There is a diference between an automatic telescope at a close by site
and one that is unattended. At a site that is automatic but also close
by
my 200VA dumb UPS is good enough. The system will see the "AC Failed"
bit and have a few minutes in which to close the current FITS file,
write a
log entry, send an e-mail notice and broadcast a message to any remote
users
then go into "zombe mode" until the power realy does go out. All this
for
a $67.00 UPS and a junk box modem.
We were also talking about a fail-safe system for a remote TASS camera
one that as Tom put it would require a couple hours drive to get to.
He wanted to be sure that his design did not preclude such a remote
installation. Many of the precautions discussed are overkill for
a system located at your house or office but if something is
cheap and easy why not do it even for a local installation?
IMO remote operations is an issue that need not be solved today.
Tom wanted to build in hooks. My opinion is that hooks are not
needed. When the time comes to set up a remote site the equipment
is available off the self.
For this envisioned remote site, one would not want to make
a four hour round trip every time the
power failed or the software crashed. We can expect both to happen.
For this kind of system it is worth building it automatic recovery from
comon failure modes. We could (at great expence and effort) build in
safty systems like those on most modern spacecraft but it is not worth
it as all we have at risk is a long drive not a $1E9 system. There
are some cheap and easy things we can do that will cover these common
problems that do not involve special built hardwere.
Further, I think we only need to design in the "hooks" so a system
_could_ be deployed to a remote location. I think it will be a
long time before anyone actualy does install a system at a remote
unattended site. My comments about the UPS
and the hardware watchdog board were just to say that there is no need
even to design in many of these hooks as they can be bough off the self
when the need come up.
For example, IMO it is a waste of effort to build in an UPS test feature
that uses an SCR to down the AC supply and force the UPS to batery mode.
Modern UPSs have this feature built-in. They go even further by
monitoring
voltages during the test.
--
--Chris Albertson
chris@topdog.logicon.com Voice: 818-351-0089 X127
Logicon RDA, Pasadena California Fax: 818-351-0699