Index: [Article Count Order] [Thread]

Date:  Wed, 21 Jun 2006 14:03:28 +0100
From:  "Colin Jack" <colin (at mark) mainline.co.uk>
Subject:  [coba-e:05827] Re: Redundancy
To:  <coba-e (at mark) bluequartz.org>
Message-Id:  <F07DD4D0940AFF41A207BE85479D2AFF180DCD (at mark) server.mainline.local>
X-Mail-Count: 05827

Load balancing is not the problem ... the data needs to sync ... and
these are mail servers, which makes it interesting ;) 

> -----Original Message-----
> From: Dogsbody [mailto:dan (at mark) dogsbody.org] 
> Sent: 21 June 2006 12:55
> To: coba-e (at mark) bluequartz.org
> Subject: [coba-e:05825] Re: Redundancy
> 
> 
> > In theory you could use a script bring down cce gracefully, tar the 
> > essential stuff and copy somewhere at like midnight or so and then 
> > bring it back up and copy to an alternate location.  There 
> are a lot 
> > of file to be concern, but should be possible. In the event 
> of a crash 
> > you will need to untar on top of an existing server and 
> will recreate 
> > the environment. cce will take a few seconds to copy so the server 
> > should not be down for more than maybe 10 seconds or so, 
> the rest is 
> > simple copy of files.
> 
> Or you just move the hard drive into a new (identical) 
> working server which would probably take just as long :-)
> 
> Surely if you get to this stage then you may as well invest 
> is some proper dedicated high-availability servers with load 
> balancers? for each of your services!?
> 
> Just my 0.02 GBP :-)
> 
> Dan
> 
>