Michael Leong wrote:
I'm doing a daily crunch of my webserver logs along w/ cache files.
However, my cache file is getting very big (700MB) and the system is
running of memory. I'm already using the LOWMEM options and turned
off unnecessary reports.
Is it possible to compact the cache file? (eg: just keep the top 100
referal urls, not every one of from the last 2 months).
You can't do that. It would lead to invalid results.
If you kept the top 100 referral URLs and http://example.com/ varies
between 98 and 105, sometimes it would not be listed when it dropped
below 100. Assume that the requests range between 9,000 and 10,000. You
could easily end up making a "combined" report from several cache files
that shows 20,000 requests for the referrer because it was only listed
in two cache files when in reality it had 40,000 - 50,000 in a one-week
period.
You can run analog several times with all but one of the LOWMEM settings
set to 3 and then take the related reports and carefully paste the
results back together. Of course this means up to six passes of analog,
but will use less memory.
Of course you could also add more memory to the system. It's getting
pretty cheap these days. ;-)
--
Jeremy Wadsack
Seven Simple Machines
+------------------------------------------------------------------------
| TO UNSUBSCRIBE from this list:
| http://lists.meer.net/mailman/listinfo/analog-help
|
| Analog Documentation: http://analog.cx/docs/Readme.html
| List archives: http://www.analog.cx/docs/mailing.html#listarchives
| Usenet version: news://news.gmane.org/gmane.comp.web.analog.general
+------------------------------------------------------------------------
