[Fwd: [squid-users] while rotating log files DNSSERVER becomes defunct (part 2)]

Hi, Henrik

I am running Squid 2.4.STABLE4 on RedHat 7.2. Here I found interesting
thing.
My /etc/resolve.conf is pointing to 192.168.1.9 DNS server. I started
squid with
DNSSERVER processes. I let it run for 1-2 hrs.
then I changed /etc/resolve.conf from 192.168.1.9 to 209.81.59.2. I did
not restart
squid and then If I say squid -k rotate and I got bunch of defunct
processes.

It might give you hint. Any help or hint or suggestion really
appriciated..

Thanks
Jigar

attached mail follows:

Hi, Henrik

Thanks for yr reply. I do see this problem after when I compied squid
with
--disable-internal-dns option.

Just for curiosity, why does squid restart DNSSERVER while rotating ?

Thanks
Jigar

Henrik Nordstrom wrote:
>
> On Thursday 05 September 2002 23.10, Jigar Rasalawala wrote:
>
> > If the DNSSERVER pending queue is overloaded by some how ( because
> > of DNS server is down for a while,
> > or because of not running enoguh no. of DNSSERVER processes ), and
> > I try to rotate log file,
> > It might cause the problem.
> >
> > Does it seem logical ? Any hint or suggestion appriciated..
>
> Should not make any difference. A new set of helpers will be started,
> and the old set of helpers will terminate when finished with what
> they were doing.
>
> The code that reaps exited child processes (defunct) is fully
> independent of this or pretty much anything else in Squid..
>
> Do you always see the problem, or does it start all of a sudden?
>
> Note: once you start getting defunct processes after a rotate I would
> expect future rotates to keep leaving defunct processes.
>
> Regards
> Henrik
Received on Thu Sep 05 2002 - 17:16:02 MDT