I already had a plan to write something like that in the past and I had
some time so I wrote this store.log tool:
http://paste.ngtech.co.il/pr3kbbf4q
The tool is written in ruby and what it does is "estimating" what is in
the cache_dir now based on reading the store.log.
Since I have not spent too much time on understanding the store.log but
I had a basic idea of whats in it that seems to give some results for now.
The tool gets only one argument and it's the location of squid store.log
and reads it like the store "journal" which takes view from nothing to
what should exit now.
Each line in the store.log represents one operation and it is expected
to be logger in the order of execution.
Due to this expectation we can predict that if a certain file was
written to the disk(using SWAPOUT) and until the end of the log(which
should represent now) it was not reported to be removed(RELEASE) from
the cache it is still there but there is no guarantee that it will be
used as a cache HIT.
The tool needs more functionality to be more accurate and to display the
estimated cache_dir size.
For now running the script piping it with "wc -l"(reduce 1 line) will
give you the result of how many objects you have in all your cache_dir
on the server from the start time of the store.log.
Any suggestions and requests regarding the tool are welcome.
Eliezer
_______________________________________________
squid-users mailing list
squid-users@lists.squid-cache.org
http://lists.squid-cache.org/listinfo/squid-users
