[imp] Stuck apache processes since upgrading to IMP 4.1

Lennon Day-Reynolds lennon at reed.edu
Mon Mar 20 14:15:17 PST 2006


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Mar 20, 2006, at 5:49 AM, Tornoci Laszlo wrote:
> I see stuck apache processes (living for more than 5 minutes and
> consuming CPU) since I upgraded from IMP 4.0.4 to 4.1. These occur
> rarely, but never die, consuming more and more CPU, slowing down
> everything. Restarting apache cures the problem, but this is still not
> nice. Does anyone else have this problem?

We have seen similar behavior with 4.0.X versions as well in our  
environment here, though only occasionally. The primary culprit we've  
been able to identify has been certain combinations of PHP  
accelerators and mod_php versions, so for the time being, we've  
reverted to running on top of a "vanilla" PHP 4.4.0, and suffering  
the performance consequences.

I'm actually experimenting right now with using Lighttpd and PHP/ 
FastCGI to host IMP 4.1, in the hopes of getting around this long- 
standing problem. I'd be happy to report back with the results of  
this testing, if other people are interested in seeing how such a  
platform behaves.

Lennon Day-Reynolds
System Support Specialist
Reed College


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Darwin)

iD8DBQFEHymvRtirLnfvQskRAgA6AKCUFQQxRF9SAIrQE62UqyYZsYd/owCgmfuz
M2YDV28VVTPpRjzQm92wOjc=
=oVI4
-----END PGP SIGNATURE-----


More information about the imp mailing list