I have one domain that has domlogs of about 100mb at this time of month does 800mb to 1gb of bandwidth a month, when analog is run via logrunner it uses 70% plus memory (256 ram, 530mb swap)


when logrunner starts, it runs through all domains setting stats for

analog
awstats
weblazer


But when it reaches this one domain and starts on analog it never seems to finish, 4hrs i waited before killing it, in fact i have to kill it every day.


any ideas

We have run analog against a user's web log that is almost 1GB and it has never just locked up. Usually doesn't take longer than 5 minutes even.

Are you using the DNS cache (looking up names for IP addresses upon analysis)?

You might try running an strace.

sorry forgot about this post :o

How can i check if dns cache setting is on?

Not familier with strace (still learning) how best to use it on analog process?


many thanks

Regarding DNS cache:

http://www.analog.cx/docs/dns.html

I would normally run analog with strace like this:

strace analog (analog options)

But since you're doing it from the control panel, you'll probably want to start it running and then from a shell, start strace on the process.

You would do a ps to find the pid of the analog instance, but you'd need to do it before the problem actually occured, otherwise you might not get any output.

ps -ef |grep analog

Then start strace on it:

strace -p pid

You might need to install strace if you don't have it installed.

http://sourceforge.net/projects/strace/