Joachim Schipper wrote:
On Wed, Jul 11, 2007 at 12:56:27PM +0200, Marc Balmer wrote:
Otto Moerbeek wrote:
On Wed, 11 Jul 2007, Adam PAPAI wrote:
The system hangs after 1 day. It's a very strongly loaded system. Any idea
why? It seems that the system does not really handle the http requests
well.
/var/www/log/access_log snippet:
httpd in malloc(): error: recursive call
httpd in malloc(): error: recursive call
[Mon Jul 9 18:33:39 2007] [notice] child pid 24755 exit signal Abort (6)
[Mon Jul 9 18:33:39 2007] [notice] child pid 10372 exit signal Abort (6)
[Mon Jul 9 18:33:39 2007] [notice] child pid 11192 exit signal
Segmentation
fault (11)
[Mon Jul 9 18:33:39 2007] [notice] child pid 15994 exit signal
Segmentation
fault (11)
The aborts and seg faults are likely caused by php. It's mostly sloppy
programming by the php developers.
A know problem exists with PHP and anonymous mmaps in the kernel. I do
not know if this is what hits you here, but the symptons look much like it.
a busy system that runs, e.g. the horde/imp webmailer will hang within a
short time (1.5 days over here).
until there is a better solution, the workaround is quite easy: restart
httpd in a cronjob.
I'm sorry to intrude here, but is this problem new? I was not aware of
it. (And as a follow-up, can we expect it to be fixed any time soon?)
it is being worked on.
