[squid-users] Question on output of mgr:carp

From: Chris Woodfield <rekoil_at_semihuman.com>
Date: Fri, 4 Sep 2009 11:27:11 -0400

Hi,

Had a quick question re: the output of the mgr:carp squidclient query.
Below is a sample of what we're seeing on our child squid:

  Hostname Hash Multiplier Factor Actual
http-cache-1a ba95c958 1.000000 0.166667 0.169094
http-cache-1b d6cf90fc 1.000000 0.166667 0.237343
http-cache-1c 274912fc 1.000000 0.166667 0.119326
http-cache-2a 9b90fb25 1.000000 0.166667 0.180795
http-cache-2b ab507efc 1.000000 0.166667 0.123492
http-cache-2c fe8e70e5 1.000000 0.166667 0.169950

The question is - does the "Actual" column count the distribution of
individual hashes to a particular parent, or the number of queries
directed there? In order words, does this point to a potential
unevenness in the hashing entropy (which I know other CARP
implementations have solved by adding entropy to the hostnames) or
just an unevenness in the query distribution, probably due to a single
popular URL being mapped to -1b?

Thanks,

-C
Received on Fri Sep 04 2009 - 15:27:15 MDT

This archive was generated by hypermail 2.2.0 : Sat Sep 05 2009 - 12:00:02 MDT