[squid-users] Follow up on squid -k rotate problem with redirector

Hi,

Now my Squid has been running almost one week without any problem
thanks to Henrik's tip...

It was not the swap space while I have 3 times the size of the squid
process but the patch did it !

After I did make clean; make... which I didn't at the first trial.

Thanks Henrik !
Lasse Holmqvist

Henrik Nordstrom wrote:
>
> Your machine temporarily runs out of swap space. You need to have at
> least twice the size of a running Squid in swap space, maybe more.
>
> If adding more swap isn't sufficient for fixing this problem then you
> could also try this patch
> <http://www.squid-cache.org/mail-archive/squid-dev/200101/0461.html>.
> But even with this patch you must have at least the size of Squid in
> free swapspace when doing a rotate.
>
> Regards
> Henrik Nordström
> Squid Hacker
>
> Lasse Holmqvist wrote:
> >
> > Hi,
> >
> > I have used -k rotate every night without any problem with
> > earlier versions of Squid but with 2.4STABLE2 it doesn't
> > work.
> >
> > New i STABLE2 is that the redirectors are restarted by -k rotate
> > which is nice while it makes it possible to rotate the rerdirector's
> > log files.
> >
> > I use squidGuard (30 children) and after each rotate I *lose* some
> > children, in cache.log this error is repeated a number of times:
> > helperOpenServers: Starting 30 'squidGuard' processes
> > ipcCreate: fork: (12) Not enough space
> > WARNING: Cannot run '/opt/local/squidGuard/filter/squidGuard' process.
> >
> > If I kill and restart squid everything works OK until next rotate.
> >
> > Squid is running on a Sun Ultra 30 Solaris 7 768 MBytes
> > with 30 GBytes ufs or diskd cache serving ~2000 clients.
> >
> > Is this a squid or squidGuard problem ?
> >
> > Regards
> > Lasse Holmqvist
Received on Wed Oct 03 2001 - 16:01:59 MDT