High Availability squid Proxy

[Schaffrath, Rene]

Hi Lingu,

we have been using linux-ha (http://www.linux-ha.org/) on our squid and postfix cluster for several years and we love it! Compared to the rhel cluster suite it's really simple to setup and maintain. Of course even linux-ha can become somewhat complicated with that fancy stuff of the later linux-ha features, but for us it just works so pretty with a simple basic setup that we have no reason to implement more complex functions. We always keep our proxy servers up to date via rhn, so we have to reboot them occasionally (kernel updates...) - and till now we never had a problem with switching the active node etc.

Since we only use plain squid and postfix I can't say much about the squidguard database. If some kind of dump-mechanism exists, I´m sure you can sync it like any other config file. I like the idea of syncing because I think of it as a defense against logical errors. If e.g. the squid cache gets scrambled (ok, has never happened but who knows...) or you rm -rf something you simply instruct linux-ha to switch the nodes and afterwards calmly solve the problem without thousand users complaining.

So I suggest you have a look at linux-ha. The downside is no support from Red Hat. But if you have a linux-ha experienced consultant, this shouldn't be a problem.

If you are interested in some more details, just contact me.

Kind Regards,
René



> -----Original Message-----
> From: redhat-list-bounces at redhat.com
> [mailto:redhat-list-bounces at redhat.com] On Behalf Of lingu
> Sent: Tuesday, October 14, 2008 1:21 PM
> To: General Red Hat Linux discussion list
> Subject: High Availability squid Proxy
>
> Hi all,
>
> I am running squid integrated with squidguard.dansguardian,clamav
> running on single  standalone RHEL 5 server.Also running webmin for
> managing squid.This Squid is serving for 4000 clients.
>
>  Since it is serving more users i don't want to take risk by running
> single server,if there is any single point of failure will cause all
> of my users to standstill.
>
> So i am planned to go for one more  RHEL 5  server with cluster.Can
> any one suggest me how to design it either i have to go for common
> storage for storing all global files or i need to synchronize both the
> server periodically by  running instance on local disk of both the
> servers.
>
> Below is  the scenario currently running on the server
>
> squid is under  /etc and squidguard database under /var/lib and
> dansguardian config file under /etc/dansguardian and webmin is under
> /usr/local
>
> Help me in designing the setup  of high availability squid
>
> Regards,
> lingu
>
> --
> redhat-list mailing list
> unsubscribe mailto:redhat-list-request at redhat.com?subject=unsubscribe
> https://www.redhat.com/mailman/listinfo/redhat-list
>

----------------------------------------------------------------
Mundipharma Vertriebsgesellschaft mbH & Co. KG
Mundipharma Straße 6
65549 Limburg (Lahn)
Amtsgericht Limburg (Lahn) HRA 624

Persönlich haftende Gesellschafterin:
Mundipharma Verwaltungsgesellschaft mbH, Sitz in Limburg (Lahn),
Registergericht Limburg (Lahn) HRB 462
Geschäftsführer: 
Mortimer D. Sackler M.D., Raymond R. Sackler M.D.,
Richard S. Sackler M.D., Kathe A. Sackler M.D.,
Jonathan D. Sackler M.D., Mortimer David A. Sackler,
Samantha S. Sackler-Hunt, Robert B. Shapiro,
Peter Boer, William Loomis, Åke Wikström, Gunther Niederheide
----------------------------------------------------------------