Re: Redundancy

> include: System maintenance, the short delay while doing a '-k
> reconfigure', log rotation, and the periodic shutdown (to tidy up memory
> usage)..not to mention the occasional fault that causes the box to

  I reboot every friday. Prior to that and any reconfigurations
  we change the dns to point to the machine were squid was initially
  developed , wait until all connections are on the other machine
  and have full access to the main proxie , once reconfigured
  rebooted , the dns is changed back , not one user should
  notice , and if they do , they are not using the correct
  dns or they using the ip , wich is not given out by us.

  Squid's performance is most satisfactory , plus it doubles
  as a tucows mirror. The alternative is paying for a commercial
  product , that works , but takes a lot of time to set up
  and maintain , with the added advantage of "hacking" the
  code to suit your specific needs , and the convenience of
  accessing it without having to walk to the ops room and plug
  in a monitor every time. Taking the performance of linux
  in general , It's a bonus thats it's free. The service
  of linux is the envy of many commercial developers out there.

  Another quick point , is there any settings , kernel configs
  that apply to a .3 sec int delay. Because we have good ping
  response , but a lot of proxie stalls , with no apparent server load.
  I just put it down to packet loss.

  Many regards
  

  Ps. Spread the gospel , register at counter.li.org (Linux user/
      machine count. 53000 so far since nov 7)

--Daniel Schroder
  Networld
  http://livewire.new.co.za (F) 419 3212
  mailto:daniel@new.co.za (T) 419 4430
  #linux (IRC) see www.zanet.org.za
  sms/mobile +2782 950 7747

On Fri, 7 Nov 1997, Scott Donovan wrote:

> >> Lately, I've been encountering a loss-of-confidence in the proxying
> >> system whenever it becomes unavailable for any reason. Such reasons
> >> include: System maintenance, the short delay while doing a '-k
> >> reconfigure', log rotation, and the periodic shutdown (to tidy up memory
> >> usage)..not to mention the occasional fault that causes the box to
> >> become unavailable (particularly infrequent, but such things _do_
> >> happen).
>
> Honestly, If you are using squid in a production environment, if possible
> apply appropriate rules. Run 2 squids as peers as the top of your
> hierarchy, and ensure as complete a coverage as possible of an proxy.pac
> that will allow load ballancing and redundancy, that way you can do your
> reconfigs/rolling upgrades etc.
>
> A hack is great, but it doesn't solve the "problem"..
>
Received on Fri Nov 07 1997 - 00:25:12 MST