[SQU] HELP with Cache selection algorithm problem

I don't understand the cache selection algorithm.

The FAQ says that the cache is selected as (more or less) the one "least
full". I assume this happens for each new request to be stored. If this is
true, how come we see a very uneven distribution? We use 4 cache partitions
and we are likely to see cache0 99% full while cache2 is 37% full. Can
someone help? We are having problems sizing these partitions correctly and
this might be related.

Thanks!

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html