[squid-users] a lot of TCP_SWAPFAIL_MISS/200

[Amos Jeffries]

On 20/07/2015 2:24 a.m., HackXBack wrote:
> top shows only 1 worker for squid ..
> 

Top shows processes. worker is a type of process, but not the only type
for Squid.

There should be at least 2 Squid processes in most installations
(master/coordinator + one or more workers). Only MacOS, Ubuntu, and
systemd machines are expected to have exactly 1 process due to their use
of -N.

If you are running one of those systems then it is more likely you will
see cache corrupted files after a Squid shutdown. Since they make
assumptions about the child process they are controlling, use abrupt
kill signals and do not always wait for the Squid to terminate cleanly.


Which goes back to how the SWAPFAIL is part of Squid self-correcting
mechanisms. Dont be panic'd by it.

Sure there is some problem/cause ... BUT Squid has already dealt with
all that. Your part is a low-pressure job of efficiency tuning -
identifying if it can be avoided in future. If not, no worries, next
time will be dealt with the same as this was.

Amos