OK,
I think I know what you are talking about Amos(hope you slept well).
The above line can cause this specific issue.
And since the helper is using the squid sources the debug output I think
looks like it's coming from within squid.
Again I might be wrong and since I do not know the squid internals of
the helpers code I am counting on Amos of knowing it.
Amos are we talking about TCP as a wrapper or as part of the code inside
the helper?
If there is a pointer towards the code inside squid I will be more then
happy to get a reference.
I do know that for a very long time Linux kernel favors TCP connections
and many programs do not like the unix pipes since they have limitations
such as one way pipe\message at at time and there for full "blocking" of
the "pipe" causing lots of stuff which is way above my current
understanding of the kernel internals.
Andrey what about the test that I have sent you with the ruby script??
I think there should be somewhere a bash fake helper but instead of
"fake helper" it should be named "debug helper" that helps the
understanding of the external_acl or any other helper on the way.
the basic mode can be OK or ERROR which is based on a simple test of 1\0
value of an argument.
Logging all input to a tmp log file...
It can be written using perl\ruby\bash\python\tcl\ash\ada\ksh\lua\php or
any scripting\shell language known in squid-users\dev.
An example in any language for "reading a STDIN line and then send
"OK\n" string to STDOUT" is welcome.
(I have seen couple lua and tcl examples here and there.)
Eliezer
On 11/14/2013 08:33 PM, Andrey wrote:
> I am curious too...
>
>
>
> --
> View this message in context: http://squid-web-proxy-cache.1019090.n4.nabble.com/Ubuntu-Server-13-10-Squid-3-3-8-WARNING-external-ACL-memberof-queue-overload-tp4663243p4663314.html
> Sent from the Squid - Users mailing list archive at Nabble.com.
>
Received on Thu Nov 14 2013 - 20:12:24 MST
This archive was generated by hypermail 2.2.0 : Fri Nov 15 2013 - 12:00:04 MST