Somewhere around Thu, 04 Apr 2002 20:48:18 +0200, Allan was seen engraving
on a handy slab of granite:
Quote:> Hello everyone,
> I just installed a SQUID (2.4-stable1) in our group Linux PC which is
> also a DNS client(I just did slightly modification).
Confusing. EVERY computer that is in any way connected to the Internet is
a DNS client; you should not need to reconfigure anything other than
putting the proper IP addresses in /etc/resolv.conf.
Quote:>Everything looks
> fine until yesterday our DNS server was down .Nobody can use netscape or
> IE to access website.
*What* do you mean here ?
a) the DNS "client" you were referring to earlier is actually the DNS
*server* that went down, or
b) the "client" could not access the DNS server and so could not access
any web sites (which is logical but can be overcome), or
c) the websites you mention are yours, and nobody in the world could
access them while the DNS server was down, which is pretty obvious...
I'm betting on a combination of a) and c), but it's anybody's guess...
Quote:> My question is :
> from SQUID doc, it said that SQUID can cache webpage which were visited
> before. but why we can not visit these webpage which were
> accessed.(these webpages should be saved by SQUID,right?)Is there any
> configuration that I missed or another reason?
There are a *lot* of configuration options for Squid - have you read the
whole squid.conf file ?
Three possibilities off the top of my head :
1. Almost ALL (popular) websites use *some* kind of dynamic content these
days, so that means web access is necessary even with Squid running as a
cache - but I'm sure you can force it to cache dynamic content as well..
2. The web site, NOT Squid, decides how long a page should be cached - if
the expiration on a page is set to 3 seconds it will always be re-fetched!
But the same as for (1) should apply here (I think).
3. You have configured Squid to use lookups (especially reverse lookups!)
for everything - even checking to see if the cached page is the most
current; this would effectively mandate the 24/7 availability of a name server
But Squid uses its own DNS lookup daemon, unlinkd, which is not dependent
on any locally running nameserver - as I said earlier, the only thing you
should need to configure is your DNS addresses in /etc/resolv.conf.
Quote:> Thank you very much!
Well, HTH !
BTW, if you have successfully installed *a* squid in a network, what I
wan to know is : where did you put the other 9 tentacles ?
--
Confusion is my middle finger.
by