Re: hierarchy problems

From: Gergely Madarasz <gorgo@dont-contact.us>
Date: Wed, 05 Jan 2000 23:54:22 +0100 (MET)

On Wed, 5 Jan 2000, Henrik Nordstrom wrote:

> > Thanks, this seems to be useful. Now I still have some problems though...
> > The ICP to the parents sometimes goes down, I don't know why. The network
> > connection is ok. And if the ICP is ok, my cache only accesses the first
> > parent listed in the config file. Now i've put no-query roundrobin to the
> > parent options, this way the traffic is really divided, but when one of
> > the parents fail (unreachable, etc), it goes to direct. Can't this be
> > somehow made to go to the other parent first?
>
> No, not without changing the code. Squid only selects one peer per
> request by default.

Hm... I thought this was like this... :/

> It should however start using the next peer as soon as Squid has
> declared the first one dead (may take a few requests, there is a
> threshold of 10 failures in a row before a peer is declared as dead)

The problem is that there are two leased lines to two different
providers... and I want to share the load. The usage of these lines is
quite often above 90% when I have the parents round-robin-ed. But this
does not provide enough security if one of the leased lines goes down...

So what I'd need here is: discover why ICP fails on both hosts at the same
time (one is two hops away, the other three, and while using them
round-robin-ed, there is no problem at all), fix this, and then configure
ICP and have my cache use both parents when available for load sharing...

-- 
Madarasz Gergely           gorgo@sztaki.hu           gorgo@linux.rulez.org
     It's practically impossible to look at a penguin and feel angry.
         Egy pingvinre gyakorlatilag lehetetlen haragosan nezni.
                   HuLUG: http://mlf.linux.rulez.org/
Received on Wed Jan 05 2000 - 16:01:38 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:50:14 MST