Apache error log fills up in no time

r3b0ot · 04-13-2015, 08:27 AM

Hi,

Setup:

Debian 6 VM
Apache/MySQL installation

since this morning my / disk fills up in no time due the error.log file in /var/log/apache2 becoming massive.

File size goes from 0 to 3.5GB in a few minutes. When I look at the log file, it's filled with thousands of times the following line:

Code:

select() error: Invalid argument

I have no idea where it's coming from. It's happening on random moments.

My only current fix is to delete the error.log file and restart apache.

any ideas?

veerain · 04-13-2015, 08:49 AM

Use logrotate to limit log file sizes and numbers.

Anyway real solution would be to find the real cause after reading (starting messages) logs.

TenTenths · 04-13-2015, 09:24 AM

passing the logs through cronolog is a better option than logrotate, as logrotate (generally) only runs once a day.

r3b0ot · 04-13-2015, 09:28 AM

Thanks.

But yeah, I'd like to find the root cause but not sure where to start.

TenTenths · 04-13-2015, 09:30 AM

So what's running under apache on the server?

Check your access_log for the same time period and see what's being recorded.

r3b0ot · 04-13-2015, 10:05 AM

Server is running about 100 dev websites. but each vhost has it's own error and access log.
so the apache error.log captures all the rest which is not vhost related

TenTenths · 04-13-2015, 10:09 AM

Disable all vhosts and see what happens.
Check you access logs to see if anyone is doing anything directly by server IP rather than hostname.
Check that whoever is developing stuff isn't using http://localhost/ anywhere. (You could stop apache from listening on 127.0.0.1 to see if this helps.