--- Forwarded Message from Christopher Jones <[log in to unmask]> ---

>Date: Fri, 23 Feb 2001 11:30:32 -0500
>From: Christopher Jones <[log in to unmask]>
>To: Language Learning and Technology International Information Forum    <[log in to unmask]>
>Subject: Re: #5967 backup for servers
>In-Reply-To: <[log in to unmask]>
>Originator-Info: login-token=Mulberry:01mgxtFf123IPRiryM5YzkhO1p5EEgd/rIb9qKmnzgng==; [log in to unmask]

------------------
Ruth:

I share your concerns, along with many others on this list, no doubt. One
cheap way of dealing with server crashes when no personnel are present is a
combination hardware/software device called Kick-off which reboots the
machine on system or application (ie WebStar) crash. We have it on our main
G4 webserver, but similar devices exist for other platforms. This helps
solve the problem of Saturday night crashes when students are reviewing
Online materials for a Monday exam.

Doesn't solve anything if the server is really dead, however.   We're a ways
away from being able to afford full redundancy for our 4 servers. We may
never go there; just assume that our students are up to waiting 24 hours
while we build a new machine from backup. In reality we should be replacing
our servers well within the MTBF (mean time between failure) predictions,
and that might actually be better policy (ie replace the production servers
more frequently, rather than doubling their number and replacing them less
frequently).

The other thing that happens on our campus is that Computing Services
offers to take things over. Streaming media is one example. We had the
first OS X server, but now they have one, they want us to move our media to
them, for a small fee...

-Chris

--On Fri, Feb 23, 2001 10:18 +0000 LLTI-Editor <[log in to unmask]>
wrote:

> --- Forwarded Message from Ruth Trometer <[log in to unmask]> ---
> 
>> Date: Wed, 21 Feb 2001 14:54:18 -0500
>> To: [log in to unmask]
>> From: Ruth Trometer <[log in to unmask]>
>> Subject: backup for servers
> 
> Greetings on a cold day in Feb!
> 
> As the user base of our servers expands, I'm becoming more concerned about
> redundancy.  I would be very happy to hear what others are doing to
> provide a fall-back plan if their server goes down - i.e. a list of
> equipment / resources dedicated to making sure all software will be
> continuously distributed.  I would also like to hear of alternatives you
> have discovered to providing 365/7/24 hour service on your own.  If you
> want to reply off list, I'll be happy to compile & repost.
> 
> Many thanks,
> 
> Ruth Trometer
> 
> P.S.  I thought there was a discussion on this topic not too long ago, but
> I couldn't find anything in the archives.  Does anyone have a reference?



Christopher M. Jones, PhD
Principal Lecturer in French / Director, Modern Language Resource Center
mail: Baker Hall 160, Carnegie Mellon University, Pittsburgh, PA 15213
tel: 412-268-5065  fax: 412-268-1328  e-mail: [log in to unmask]