On 17/05/11 22:45, gaël therond wrote:
> Dear SQUID Support team,
>
> I'm facing a quite strange problem today, I'm running a SQUID 3.1.12.1 RC and I have to increase the amount of FD available for SQUID, due to the amount of clients (over 1000).
>
> I've change my sysctl.conf to be able to use 256 000 FD on my machine . I've add the following line on the sysctl.conf file:
> fs.file-max = 256000
>
> Then now, I've stop the squid service, erase the /usr/local/squid folder, erase the /cache01/ partition (cache file partition).
>
> Re-compile squid with the following directive:
>
> ./configure --prefix=/usr/local/squid/
> --sysconfdir=/etc/squid/ --enable-async-io --enable-snmp --enable-icmp --enable-wccpv2 --enable-kill-parent-hack
> --enable-cachemgr-hostname=SQUID --enable-htcp --enable-cache-digests --enable-ssl --enable-linux-netfilter
> --with-openssl=/usr/lib --with-large-files --with-filedescriptors=256000
>
> and of course run mak and make install (as root).
>
> BUT I still have the following message on my squid.log file:
>
> With 1024 file descriptors available
>
> Shouldn't be With 256000 file descriptors available?
3.1 supports the max_fd config option, no need to recompile anymore :).
There is some bug as yet untraced breaking the max_filedescriptors
squid.conf option from working without ulimit.
For now you still need to change squid.conf and run ulimit before
start/restart/reconfigure to raise the number Squid sees available from
the OS.
Amos
-- Please be using Current Stable Squid 2.7.STABLE9 or 3.1.12 Beta testers wanted for 3.2.0.7 and 3.1.12.1Received on Tue May 17 2011 - 11:59:06 MDT
This archive was generated by hypermail 2.2.0 : Fri May 20 2011 - 12:00:02 MDT