[asterisk-users] asterisk at 100% CPU, 1000's of log files

[Tony Mountifield]

In article <094A73044298734FB7D58CAAA319E1D66AC576 at UBIQ-SERV1.ubiquisys.local>,
Adrian Marsh <Adrian.Marsh at ubiquisys.com> wrote:
> Hi All,
> 
> Twice now in the past few weeks I've walked into the office to find that
> our 1.2.24 Asterisk process is sat at 100%, and that hundreds of
> thousands of log files in /var/log/asterisk exist, all at 312 bytes,
> containing:
> 
> Aug 29 23:22:17 VERBOSE[24303] logger.c: Asterisk Event Logger restarted
> Aug 29 23:22:17 VERBOSE[24303] logger.c: Asterisk Queue Logger restarted
> Aug 29 23:22:17 VERBOSE[24303] logger.c: Rotated Logs Per SIGXFSZ
> (Exceeded file size limit)
> Aug 29 23:22:17 WARNING[24303] format_wav.c: Unable to find our position
> 
> Stopping/starting A*k seems to reset things ok...

Hi Adrian,

This is a problem I experienced and reported at http://bugs.digium.com/view.php?id=8360

We never did agree on the best way to overcome it, and the bug got
closed/suspended when I didn't have time to follow up on it.

My guess is that you had a recording running somewhere whose file size hit
2GB. On a 32-bit system, this generates a SIGXFSZ, which prompts Asterisk
to rotate the logfiles (without actually checking whether it really was a
logfile that caused the SIGXFSZ). Since it was caused by something else,
the signal keeps getting generated and the logfiles keep getting rotated.

On my system, it was Meetme recordings that could cause it. I sidestepped
the issue by putting a filesize check into Meetme before writing a frame
out to the record file.

Cheers
Tony

-- 
Tony Mountifield
Work: tony at softins.co.uk - http://www.softins.co.uk
Play: tony at mountifield.org - http://tony.mountifield.org