Hmm. What OS are you running? I think you said before that it was Linux,
but I can't remember for sure. I have seen the crashing problem before on
BSDI machines, and managed to fix it by tweaking the cache_mem settings
*very* carefully.
Provide a bit more info about what happened when your cache crashed and
how it is configured.
And I didn't even notice your cache was down... Hmmm. Must have happened
when my cache was querying some other cache :-)
        
        -Bill
On Fri, 26 Sep 1997, Abha Ahuja wrote:
> 
> 
> On Fri, 26 Sep 1997, Michael Pelletier wrote:
> 
> > On Thu, 25 Sep 1997, Bill Wichers wrote:
> > 
> > > Turn source_ping on. source_ping lets squid include the original site
> > > along with any parents/siblings when it decides where to fetch an object
> > > from. source_ping is off by default.
> > 
> > Hm, interesting.  I thought that if it could determine that the parent was
> > down that it would just revert to the original site unless you had
> > firewall options set.  Thanks for the tip, hopefully that'll do the trick.
> > A better solution of course would be to have the upstream cache not
> > crashing, but oh well...  :-)
> 
> Hi All!
> 
> I'm the upstream cache that crashed... Anyone ever see this behavior
> before?
> 
> Our cache was happily chugging along when all of a sudden for no apparent
> reason, the squid process hung.  Giving it a HUP got everything going
> again, but we aren't sure what caused the problem and how to avoid it in
> the future... We are running squid-1.1.15 plus the connect-retry patch on
> a linux box.
> 
> Anyone have any ideas?
> 
> -abha ;)
> 
> 
> __________________________________________________________________________
> --------------------------------------------------------------------------
> abha ahuja						ahuja@merit.edu
> Merit Network, Inc.					313.764.0294
> 
> 
Received on Sat Sep 27 1997 - 10:55:25 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:37:11 MST