> Wow. Sure thats hits/sec and not hits/minute?
> The 'extreme' setups of Squid-2.7 only reached 990req/sec.
I'm running squid3.0 on Dell R300 servers with 4x2.8GHz Intel Xeons and 12GB of ram. On production servers I'm getting
max 1500hits/s. With 2500hits/s I have seen that some in access.log, in elapsed column that some requests were closed
after 6 seconds and average was something like ~300ms. It's not acceptable for me. Most requests on production servers
are closed in less than 1ms.
>FWIW; the only other occurrence of this particular "Select loop Error" reported in recent years was found to be due to
broken NIC drivers.
>The behaviour sounds very much like some such bug has been hit, or maybe a limit on the open ports per IP.
I will search for something about nic drivers on FreeBSD maillists. As I said before squid3.0 was running fine for few
hours, everytime I launched squid3.1 it lasted only few seconds. There is no firewall configured on this machine.
Could anyone give me some pointers about debugging newest squid to see what is causing that squid immediately closes
connections?
Received on Mon May 24 2010 - 09:32:00 MDT
This archive was generated by hypermail 2.2.0 : Mon May 24 2010 - 12:00:05 MDT