Re: hierarchy problems

From: Gergely Madarasz <gorgo@dont-contact.us>
Date: Thu, 06 Jan 2000 22:07:04 +0100 (MET)

On Thu, 6 Jan 2000, Henrik Nordstrom wrote:

> Patches of mine I'd recommend you to apply (in this order, or with
> --fuzz=3, in which case the icp_dead_detection patch can be skipped if
> so wanted):
> squid-2.2.DEVEL3.icp_dead_detection.patch
> squid-2.2.STABLE3.tcp_dead_detection_and_retry.patch
> squid-2.2.STABLE2.peer_connect_timeout-2.patch
>
> > How will it detect when the dead peer comes back?
>
> It will probe the dead peer with TCP connections until it answers again.

I've just tried this out, it seems it really works, thanks :)
But don't think I'm finished with my questions :)
So the following seems a bit problematic to me:
peer A fails, my proxy fetches everything from peer B for an hour...
the fetches count for peer B is much higher. peer A comes back, and for
the next hour or so my proxy wants to fetch everything from peer
A to even out the fetched counts for both peers. This again works against
sharing the load on the leased lines evenly... couldn't squid be told not
to check for the fetches count since the beginning but for the last 5
minutes or something in the round-robin algorithm ?

-- 
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 Thu Jan 06 2000 - 19:14:56 MST

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