Re: [squid-users] squid TCP_MISS/503 and "(105) No buffer space available"

From: Eliezer Croitoru <>
Date: Mon, 21 Oct 2013 11:33:35 +0300

Hey there,

There are couple unknowns in the picture:
What vesrions of squid are you using? "squid -v"
What is the network size?
What is the server load?
What is the machine properties?
You can use the cachemgr interface to get some information out of squid
to understand the situation.


On 21/10/13 10:31, monte wrote:
> Hello,
> The squid is running good, but after some time it begins getting
> TCP_MISS/503 in access.log and "(105) No buffer space available" in
> cache.log for all sites and all users in browsers begins getting "ERROR The
> requested URL could not be retrieved". After some minutes everything get
> back and work good until next time. Where can be problem?
> Also I had configured transparent proxy using iptables:
> iptables -t nat -A PREROUTING -i eth1 -p tcp --dport 80 -j DNAT --to
> eth1 - is my LAN
> - is squid server in localhost
> CPU usage is max 24%, RAM usage about 50%.
> I have tried disable caching to slow down RAM usage, but the same problem.
> *In old rc.local config I found this line "#ulimit -HSn 8192", perhaps it
> can be the solution for this? Need I recompile squid to work with this
> command?*
> --
> View this message in context:
> Sent from the Squid - Users mailing list archive at
Received on Mon Oct 21 2013 - 08:33:46 MDT

This archive was generated by hypermail 2.2.0 : Mon Oct 21 2013 - 12:00:06 MDT