[Tutor] Logfile multiplexing
Alan Gauld
alan.gauld at btinternet.com
Wed Nov 11 11:05:09 CET 2009
"Stephen Nelson-Smith" <sanelson at gmail.com> wrote
> I don't really want to admit defeat and have a cron job sort the logs
> before entry. Anyone got any other ideas?
Why would that be admitting defeat?
Its normal when processing large data volumes to break the
process into discrete steps that can be done in bulk and
in parallel - thats how mainframes have worked for years.
Its a perfectly valid approach to preprocess your input data
so that your main processing can be more efficient. The trick
is to spread the load where the task is repeatable (eg sorting
a file - and even the filteriing of your php's) and maximise
the efficiency where it is not (ie merging).
So it would be valid to have a set of batch jobs removing
the phps, followed by another job to sort each file, then finally
merge the reduced files. The initial filtering and sorting can
both be done on a per file basis in parallel.
HTH,
--
Alan Gauld
Author of the Learn to Program web site
http://www.alan-g.me.uk/
More information about the Tutor
mailing list