Re: [squid-users] What load can SQUID cope with in accelerator

From: Robert Collins <>
Date: Fri, 16 Mar 2001 20:47:07 +1100

Possibly buffered I/O. If you haven't turned off buffered I/O in your
redirector, do so..


----- Original Message -----
From: "Ole Moller" <>
To: "Brett Lymn" <>
Cc: "Duane Wessels" <>; "Simon Morley"
<>; <>
Sent: Friday, March 16, 2001 8:12 PM
Subject: Re: [squid-users] What load can SQUID cope with in accelerator

At 00:24 16-03-2001, Brett Lymn wrote:
>According to Ole Moller:
> >
> >(instead of expected 1mb/sec). When squid had run for about 10
minutes it=20
> >killed itself with a message of "FATAL: Too many queued redirector=
> > requests".
> >
>Sounds like you need to have more redirector children - the log
>messages say you should do this. If you are still having troubles
>then you should look at the coding of your redirector and try to make
>it more efficient so they do their job quicker.

I have the maximum of 30 configured, please remember that when using the
exact same script whithin apache 1 was sufficient. Further more:

# time head -1000 access.log | awk '{printf("%s\n",$7)}' |
../bin/ > /dev/null
         0.06 real 0.00 user 0.00 sys

At a peak of 20 req/s and 30 redirectors to handle it, I dont think the
real problem is the redirector but "something" else.

> >I believe that this not is the real cause of death considering that I
> >started 30 and one did the job fine when running from apache. Have
you any=
> >=20
> >idea of what could be blocking? disk io?
> >
>What does iostat & vmstat say? These should give you some clue as to
>where the bottleneck is.

There is disc activity, but the numbers indicate that they only are
about 60% of maximum.


Ole Møller, Sysadm CyberCity Internet
Received on Fri Mar 16 2001 - 02:47:41 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:58:40 MST