Problem with squid/socks5 ?

From: Michael O'Sullivan <mos@dont-contact.us>
Date: Fri, 11 Apr 1997 09:55:23 +1000 (EST)

HI all,

I'm new to the list and have just taken over running a squid proxy, so excuse
me if my questions are rather simple. I've had a quick look in the FAQ and src
doco I have found for an answer.

Basic situation is squid running behind firewall using socks5-beta-0.16.6 to
get out to internet. I was running 1.1.1, but upgraded to 1.1.9 lat night,
however probelm persists.

Squid will run along fine for a while, and the go CPU bound. Looking at the
cache.log file it seems that once it gets a there are a number of

ERR_CLIENT_ABORT,

around the time that the process goes off into the loop.

a truss of the process shows
......
poll(0xEFFFC610, 1, 15000) = 1
read(19, 0xEFFFE990, 2) = 0
fcntl(19, F_GETFL, 0x00000000) = 6
ioctl(19, I_FIND, "sockmod") = 1
fcntl(19, F_GETFL, 0x00000000) = 6
ioctl(19, I_FIND, "sockmod") = 1
poll(0xEFFFC610, 1, 15000) = 1
read(19, 0xEFFFE990, 2) = 0
fcntl(19, F_GETFL, 0x00000000) = 6
ioctl(19, I_FIND, "sockmod") = 1
fcntl(19, F_GETFL, 0x00000000) = 6
ioctl(19, I_FIND, "sockmod") = 1
poll(0xEFFFC610, 1, 15000) = 1

........
The squid process goes to 100% CPU bound, and basically must be killed and
restarted for anything to work again. I'm in the process of looking into the
socks side of things (whihc is where I think the problem is), but was wondering
if anyone out there has already seen this problem and can point me to a fix.

I'll keep the list informed of any fixes I find myself,

Thanks in advance,

Michael
Received on Thu Apr 10 1997 - 16:58:52 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:34:58 MST