On Tue, 28 Jun 2011 13:17:56 -0300, Usuário do Sistema wrote:
Hello everyone, sometimes my squid process go to down and quickly to up again.maybe less of 3 seconds.
Yes. An automatic restart after the crash.
in the logs ( /var/log/messages ) appear follow message: Jun 27 07:23:11 jvelnxfw02 squid[25247]: Too many queued ntlmauthenticator requests (26 on 5) Jun 27 07:23:11 jvelnxfw02 squid[25245]: Squid Parent: child process 25247 exited due to signal 6 Jun 27 07:23:14 jvelnxfw02 squid[25245]: Squid Parent: child process 26127 started I'm using the Active Directory to authentication. follow my squid parameters about ntlm auth_param ntlm program /usr/bin/ntlm_auth --helper-protocol=squid-2.5-ntlmssp auth_param basic children 20 auth_param basic realm Acesso a Internet Tigre SA auth_param basic credentialsttl 2 hours
"basic" authentication protocol settings have no effect on "ntlm" authentication protocol.
What this means is that you are using the default value of "auth_param ntlm children 5".
The message Squid announced also said "(26 on 5)", meaning the configured _5_ ntlm helpers were trying to handle 26 requests simultaneously. They can handle at most one, and have one queued waiting. So a total capacity of 10 new client connections to your Squid at any one time. Modern web browsers will consume a large chunk of that low capacity (4-6 of the 10) loading a single web page.
Amos
