Apache2 & child pids

English Network/Internet
#1

Hi,
i hope u can give me same tips. aboute apache-prefork. On my server i running an apache-prefork with some modules


actions alias auth_basic authn_file authz_host authz_groupfile authz_default authz_user authn_dbm autoindex cgi dir env expires include log_config mime negotiation setenvif ssl suexec userdir status dav dav_fs dav_svn authz_svn ldap authnz_ldap jk php5

After a while i getting following error in error_log


[Thu Oct 20 09:21:48 2008] [notice] child pid 18117 exit signal Aborted (6)
[Thu Oct 20 09:21:48 2008] [notice] child pid 18119 exit signal Aborted (6)
[Thu Oct 20 09:21:48 2008] [notice] child pid 18120 exit signal Aborted (6)
[Thu Oct 20 09:21:48 2008] [notice] child pid 18121 exit signal Aborted (6)
[Thu Oct 20 09:21:48 2008] [notice] child pid 18222 exit signal Aborted (6)

and the PIDS going to 100% cpu load and my hole system getting a load from > 30. Ldap is for the authentification with a AD-Server. If the clients surfing a page, it need login in wiht his AD-Account.

My question is, why can apache dont close the pids.

#2

What do you mean by close the pids? Your apache processes are all dying abnormally, which is why the numbers keep on changing. Something is wrong with a module you are using.

#3

I think, that apache will close a not used http-prefork process, but anythink impeded that. One the same time, one pid goes to 100% it will never be closed.

I think so too, that the problem are the mod_ldap but im not really sure!

#4

Pids are not closed in the way file descriptors are. Either the child process exits normally or it does not. When it does not, the parent apache process receives a non-zero status and reports the problem of the child exiting abnormally in the log. When the child exits without having done any work then it’s worse because the parent starts up another child to replace the dead one, and this keeps destroying and creating new processes, eating up your CPU.