NYCPHP Meetup

[Hans Zaunere]

Sexton, David wrote:

> Hey Gang,
>  
> After a few months of trying to figure out why we were getting sets of
> maximum execution timeout errors at totally random times, I've narrowed it
> down to being the result of adhoc queries that return very large datasets. I
> noticed something interesting though, and I was wondering if anyone had any
> information on this...
>  
> First of all, it's PHP build 4.3.2 running in CGI mode on Win32. We have the
> max_execution_time directive set to 90 seconds, but whenever one of these
> wide-open adhoc queries is executed, the spawned PHP process eats up over
> 90% CPU time... but for about 6 minutes in real time. This is leading me to
> believe that PHP's internal timer is actually being short-changed by it's
> own CPU intensive existence. :)  Anyway, it seems like what PHP "thinks" is
> 90 seconds is actually more like 5-10 minutes in reality. Is this a bug, a
> WIN32 specific problem, both, etc... any information would be appreciated.

I have heard things similar to this kicked around, although I don't know many specifics, unfortunately.  Win related may be an issue as well, and might surface as a bug with some finite testing.

I've seen that http://bugs.php.net reports some of these issues (search for max_execution_time).

H