Re: [squid-users] NMBCLUSTER

The first is most likely harmless, caused by a unavailable DNS domain
(no responding DNS servers). The message could be a bit more expressive
in what it gives up on...

--
Henrik Nordstrom
Squid Hacker
TomC wrote:
> 
> Hi,
> My company has 3 proxy running with PIII 800, 1G RAM, 90GB cache and
> BSDI 4.1. At start, it ran smoothy.
> But after each cache fill up over 25%, 3 servers always crash. I found
> that in the /var/log/messages
> 
> May 19 18:24:07 ngautaukok squid[291]: idnsCheckQueue: ID 45f3: giving
> up after 20 tries and 100.6 seconds
> May 19 21:01:01 ngautaukok kernel: mb_map full, NMBCLUSTERS (8192) too
> small?
> 
> I know that i need to increase the size of NMBCLUSTERS to 32MB, but I
> don't know what is the meaning of first
> log? Thx
> 
> Brian