we have many sites located with an Amazon . com x-large AWS server. Whenever i restart apache, i'll run top and find out quantity of apache processes (6-8) which will rapidly increase cpu percentage and memory percentage as much as 33% cpu, sometimes 50% cpu or more to at least one.3rd generation of memory each! they'll remain in top for ten to twenty min, then another will begin up with similar behavior.
basically do an strace on one of these simple processes i'll have an edless list such as this below:
access("/etc/localtime", R_OK) = 0 access("/etc/localtime", R_OK) = 0 access("/etc/localtime", R_OK) = 0 access("/etc/localtime", R_OK) = 0 access("/etc/localtime", R_OK) = 0 ...
their email list won't stop. basically leave these processes going, they'll eventually stall the server and i have to reboot the whole virtual machine.
anybody have applying for grants why the procedure keeps being able to access this file, or any other method of the way i can discover why this really is happening?
Should you enable
mod_status and hang
ExtendedStatus on, you'll be able to visit
http://server/server-status (or any other location you place to make use of the
server-status handler) and you will get a table of the items each Apache process/thread does:
Srv PID Acc M CPU SS Req Conn Child Slot Client VHost Request ⋮ 3-0 2129 0/167/167 _ 1.13 495 0 0.0 0.33 0.33 ::1 Zia.metrics.net GET /mrtg/mrtg-l.png HTTP/1.1 4-0 2130 1/150/150 W 1.11 0 0 1.4 0.30 0.30 ::1 Zia.metrics.net GET /server-status HTTP/1.1 ⋮
The status page incorporate a full key and area description, in addition to another summary status. You need to most likely safeguard it to simply be visible for you, not the web:
<Location /server-status> SetHandler server-status Order deny,allow Deny from all Allow from 127.0.0.1 ::1 </Location> ExtendedStatus On
You should use any access control method obtainable in Apache (e.g., Fundamental or Digest authentication), obviously.