Tugger the SLUGger!SLUG Mailing List Archives

RE: [SLUG] Apache on a cluster


The second firewall is just incase 1 dies the other will take over including
the second web server.

I think the web site's code is written to access a number of servers to
handle parts of the web site so if the client asks for sydney based info,
the web server knows it's on database 3 or 4 (3 and 4 are mirrored data and
not neccessarily clustered) and requests the info from the MySQL database.

I'm not sure how it will handle requests where server 3 fails, possibly
it'll just ask server 4 if the data didn't come back properly and send an
alert to the admin somehow...

Something I'll have to ask the customer to ask their programmers on what
setup it's expecting..

thanks,
George Vieira
Systems Manager
Citadel Computer Systems P/L


-----Original Message-----
From: Kevin Saenz [mailto:kevin@xxxxxxxxxxxxxxx]
Sent: Wednesday, 21 November 2001 11:58 AM
To: slug@xxxxxxxxxxx; slug-request@xxxxxxxxxxx
Subject: Re: [SLUG] Apache on a cluster


> George Vieira wrote:
> 
> > I'm not sure if I've asked this before but can Apache run on a
> cluster
> > network? I've searched their site and don't really see any mention of
> it or
> > how to configure it, does this mean it's not cluster aware?
> > 
> > I'm trying to design/build a redundant web site which contains 3
> stages (non
> > clustered version) as below...
> > 
> > Link--+--Firewall1---Webserver1--+----+----Database1
> >       |                          |    +----Database2
> >       +--Firewall2---Webserver2--+    +----Database3
> >                                       +----Database4
> > 
I believe that if you have a look at piranah I think it can do something
Like this I am pretty sure that some of the insecurities have been removed.
internally we have developed a clusered failover cluster for our
authentication
server which is running LDAP. if one server goes down the redundant server
will in turn bring up the services that you require and act as though there
was no problem. I thing the only issue you could have is if there are
sessions connecting the databases when ever someone connections to the
website.
I think that you would need to have some form of service where both machines
will understand what the other is processing so that if one goes down the
other
will pick up the job and run with it. Your design appears that you are 
implementing the firewall on the web server?
I hope I have answered your problem


> > should the firewall,webserver or any of the database servers fail,
> the
> > other(s) would take over.
> > 
> > Though, my clients software will be doing region checking via the URL
> path
> > (sydney,perth,etc) and will direct the traffic/webrequests to the
> > appropriate Database server which handles that region....
> > If a cluster was used I don't think that would then be needed..
> > 
> > any ideas/help?
> > 
> > thanks,
> > George Vieira
> > Systems Manager
> > Citadel Computer Systems P/L
> > 
> > 
> > 
> 
> 
> 
> 
> 
> --__--__--
> 
> --
> SLUG - Sydney Linux User Group Mailing List Digest -
> http://slug.org.au/
> More Info: http://lists.slug.org.au/listinfo/slug
> 
> 
> End of slug Digest
> 



-----------------------------------------------------------------------
Kevin Saenz
Security Analyst
mobile: +61418455661
email: kevin@xxxxxxxxxxxxxxx

-- 
SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/
More Info: http://lists.slug.org.au/listinfo/slug