JSiergiej@xxxxxxxxxxxxxxxx wrote:
I am running squid 2.6 on a RedHat Enterprise box in my DMZ. If point my
public addresses to secondary non-routable IP addresses on the NIC of the
Linux box via ACL's on my firewall. From there, requests are filtered into
my internal web servers in regards to the non-routable IP the request is
coming from.
I have having an issue where my squid reverse proxy will only function
correctly if I open a terminal window and start the squid process with
/usr/local/squid/sbin/squid -NCd1. My webpages will work correctly from
this point on and I notice it starts with only one process id when I do it
this way.
Hmmm... If I'm not mistaken, RHEL installs the squid binary by default
as /usr/sbin/squid. I'd be willing to bet that's the version that the
startup script is utilizing. That Squid was compiled to use a different
conf file (/etc/squid/squid.conf). I imagine that it's logging to
/var/log/squid/*.log.
So, you have three choices. From simplest to most complex...
1) Disable the Squid service from starting ("chkconfig squid off" on the
command line is one method). Then edit /etc/rc.local and add the
"/usr/local/squid/sbin/squid -NCd1" (without quotes) as the last line.
This will start your new Squid with the server.
2) Edit /etc/init.d/squid, and replace every instance of /usr/sbin/squid
with /usr/local/squid/sbin/squid. This should use the native startup
script to manage the new Squid.
3a) Recompile Squid from source to replace the RPM version. Run
/usr/sbin/squid -v to get the compilation options of the RPM version
(the list will be long, and will have some repeated arguments). Pay
close attention to the last instance of any *dir options (bindir,
sbindir libexecdir, etc). Recompile Squid with the options you selected
feel beneficial, and the *dir options from the RPM.
3b) Grab the Fedora 6 SRPM (squid-2.6.STABLE9-1.fc6.src.rpm) and build it.
However, if I use the actual squid service in the System Tools instead of
using the command line, I see that it always starts with 2 process ids and
my webpages will not work. I get page not found and no traffic shows up in
my access.log file.
I need this service to start automatically and function because the Linux
box is rebooted after hours once every week. It rebooted last night and I
came in this morning to find that the websites were down so I ran the squid
startup command manually and everything returned to normal. I tried using
the squid service itself, which started normally even on reboot, but the
websites would not work until I shut off the squid service and ran the
startup command normally.
If there are any ideas as to how I can resolve this issue and have squid
startup automatically, I would appreciate the help. I am not an expert at
Linux so please feel free to explain things in a more detailed manner if
need be.
Hopefully the options outlined above are clear enough. If not, feel
free to ask more questions.
Thanks in advance,
Jack Siergiej, MCSA
Chris
