🚀 go-pugleaf

I would really appreciate it if someone could help with the following
problem.

I've set up and installed Squid 1.1.14 under Linux. We previously had been
running a Harvest proxy server with a 500Mb cache, and with Squid its now
3.4Gb. 

All seemed to be going well initially. However, there seems to be a LOT of
errors (seems to be about 50% of requests actually) being reported by the proxy 
server on sites that we KNOW are fine and connect fine one minute and fall over 
the next, with an error:

ERROR
The requested URL could not be retrieved

While trying to retrieve URL http://whatever

The following error was encountered:
  Connection Failed

The system returned:
  (113) No route to host

This means that:
  The remote site or server may be down. Please try again soon.

Generated by squid/1.1.14@rainforest.wolf.net.au


In the cache.log file there are corresponding entries:
 ERR_CONNECT_FAIL: http://whatever.com

I've also seen this error appear and individual images not appear on the web
page returned by the proxy server but with no full screen error.

This is also my second attempt at running this. I had to wipe the hard drive
partition the cache is on because the /cache directory seemed to be corrupt,
with files of type "?" being created (where the "d" for directory would
normally be) and owned by user 4139 or somesuch.

That doesn't seem to be happening this time, yet...

Our router has port 80 blocked, as a measure to force people to use the
proxies as well.

I know the sites are working, because when we try to access them via the old
proxy server they work fine all the time.

I had thought maybe this could have been caused by a memory problem but the
machine itself has 64Mb and I've allocated 16Mb to it in the squid.conf
file. I've also noticed the proxy server shutdown several times with this
error:

97/10/15 20:31:01| commBind: Cannot bind socket FD 17 to *:8080: (98)
Address already in use

What could be causing this? The server seems to run again immediately
because of the RunCache program which apparently will restart the server if
it dies.

Also, when it does shutdown, it shows the following memory stats which just
don't look very safe.. the amounts look a little low for a machine with 64Mb
of RAM.

Squid Cache (Version 1.1.14): Terminated abnormally.
CPU Usage: user 0 sys 0
Maximum Resident Size: 0 KB
Page faults with physical i/o: 134
Memory usage for squid via mallinfo():
        total space in arena:     393 KB
        Ordinary blocks:          343 KB      1 blks
        Small blocks:               0 KB      0 blks
        Holding blocks:           260 KB      1 blks
        Free Small blocks:          0 KB
        Free Ordinary blocks:      49 KB
        Total in use:             603 KB 154%
        Total free:                49 KB 12%

All in all this is very frustrating. The documentation on the Squid home
page seems rather sparse and doesn't go anywhere near addressing ANY of
these issues.

If anyone has ANY ideas or suggestions at all, then they would be greatly
appreciated to rainbow@zip.com.au

Regards
Douglas Garstang

-- 
----------------------------------------------------------------------
|\/\/\/| Douglas Garstang AKA Automan    B.Bus(Comp)  _--_|\
|      | 6 Bellinger Road Ruse Campbelltown NSW 2560 /      \
(o)(o) | rainbow@zip.com.au      Mobile# 017 829 027 \_.--._/<-- Here!
|c     ) dwgarstang@teletechintl.com                       v
 \--  |  -------------------------------------------------------------       
  \   |  I'm so hip I can't see over my own pelvis!  
   -------------------------------------------------------------------