> >On 21.03 16:20, Juhani Tali wrote:
> >>One way might be to add to all 1-st level squids parents to
> >>squid-office.conf
> >>cache_peer cache1 parent 3128 7 no-query no-digest default weight=2
> >>cache_peer cache2 parent 3128 7 no-query no-digest default weight=1
> >>....
> >>Will this work with no-query?
> Matus UHLAR - fantomas wrote:
> >I would not setup no-query. I would set up parent proxy for relationship
> >with each of those squid instances, based on true ICP (on different
> >ports/ips).
> >
> >the 1st-level proxy can decide who to use as parent for each request.
On 22.03 10:30, Juhani Tali wrote:
> The setup that I was handed over was without the ICP. Most probably it
> was finetuned, because the CPU-s are running at 101%
that does not mean the proxies were fine-tuned. that can indicate they
were not :)
> For me ICPs only use is that it allows the clients to load-balance based
> on response time from the servers. (all queries to all servers?) All the
> external traffic must pass through the squid-s anyway.
along with cache digests, ICP can be used to detect if one of siblings has
object in its cache. That's why I recommend using it.
> Now I am starting to think, that if the queried parent squid will ask
> with ICP from other parents/siblings anyway, then what is the real
> difference...
see http://squid.visolve.com/squid/squid24s1/glossary.htm for difference
between parent and sibling.
If you have more 2nd-level squid caches with different links, you probably
should set up them all as parents to your 1st-level proxy and as siblings
to each other (just for sure) and set up ICMP on all of them to be able to
know which proxy is closest to the requested data and fetch the data using
that parent.
-- Matus UHLAR - fantomas, uhlar@fantomas.sk ; http://www.fantomas.sk/ Warning: I wish NOT to receive e-mail advertising to this address. Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu. Linux - It's now safe to turn on your computer. Linux - Teraz mozete pocitac bez obav zapnut.Received on Tue Mar 22 2005 - 03:02:34 MST
This archive was generated by hypermail pre-2.1.9 : Fri Apr 01 2005 - 12:00:02 MST