hi!
version - Squid-2.3.STABLE4
how does the icp_query_timeout works ...
reason i am asking this is ...i tried this ...
setup a parent(freeBSD) and child(LINUX) ... set the icp_query
timeout of child to 2sec. ...introduced a delay of 3sec between them
(using dummynet on the parent) .... and send a couple of thousand reqs. to
the child ...
i got about 800 timeout_directs and around 1100 first_parent_misses ...
whereas it should have been almost all timeouts ... i assumed that the
dummynet delay is not very accurate ... i want to confirm whether my
assumption is true ... or there are some issues with the squid timeout ...
The squid.conf says that by default squid determines it dynamically based
on previous values. if its dynamic what algorithm does it uses ... can
anyone give me some information about it ... in the above case i set a
fixed timeout ... so is the dynamic part interfering with the fixed value
to give such results ... i tried the above setup 4-5 times with different
traces ... but i never got all timeout ...
Received on Thu May 31 2001 - 11:24:33 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:00:22 MST