Bug 28968

We are running Red Hat 7.2 (Kernel 2.4.7-10) with Apache 2.0.49 (using simple 
configurations, nothing fancy, compiled from the sources). This friday morning 
we had a call from our customer that their web site was down. We noticed that 
this was the case for 100 other customers too. We also noticed that this was 
because Apache wasn't running. Of course we tried to start it but it wouldn't 
start. First it was because the PID-file existed. Erasing it helped a bit - 
Apache is running again.. but it died immediteally after the first request. So 
we thought to check the semaphore arrays if they were filled up - clearing them 
didn't help either. After 30 minutes when we had tried everything.. we checked 
the error log and saw it was huge (though there was nothing intresting). We 
tried rotating it.. and voila, apache runs smooth and fine again. The error log 
was 2147483647 bytes in size. My guess is, when apache tried to write something 
in the log, it crashed somehow because it was so huge. Apache only crashed on 
pages which had missing images or anything else that caused an entry in the 
error log.