Script to determine Apache ServerLimit and MaxClients.

check-apache-instances.sh

#!/bin/sh
available=$(free --mega -tw | grep "Total:" | awk '{print $4}')
ps -ylC httpd --sort:rss | awk -v avail="$available" '{ s += $8; } END { print "Average Size:", s/(NR-1)/1024, "MB,", NR-1, "Processes, Total usage:", (s/(NR-1)/1024)*(NR-1), "MB, Max Servers:", avail/(s/(NR-1)/1024) }'

httpd.conf

<IfModule prefork.c>
    StartServers            10
    MinSpareServers         10
    MaxSpareServers         25
    ServerLimit             128
    MaxClients              128
    MaxRequestsPerChild     0
</IfModule>

shell-output.txt

$ free -m
             total       used       free     shared    buffers     cached
Mem:          7454       1405       6049          0        142        401
-/+ buffers/cache:        860       6594
Swap:         1023          0       1023

$ ./check-apache-instances.sh
Average Size: 26.043 MB, 37 servers, Total usage: 963.59 MB, Max Servers: 249.588

Hi i could not find the MaxClients 128 in the httpd.conf file in Appache xammp.
also this line to

can you please tell me what i am doing wrong..

Hi, great script. However, you'll need to account for the Header Line in the ps output. I modified your script by adding (NR-1) in the relevant locations. Thank you!

ps -ylC httpd --sort:rss | awk '{ s += $8; } END { print "Average Size:", s/(NR-1)/1024, "MB,", NR-1, "Processes, Total usage:", (s/(NR-1)/1024)*(NR-1), "MB, Max Servers:", 6500/(s/(NR-1)/1024) }'

how to get maxclient limit

I have 65 gb ram so what should be the maxclient ? i have already set maxclient to 900 but at some point it exceeds that value.

Thanks for the idea and script. I modified it a little to accept the actual free resources from the output of free -tw

#!/bin/sh
available=$(free --mega -tw | grep "Total:" | awk '{print $4}')
ps -ylC httpd --sort:rss | awk -v avail="$available" '{ s += $8; } END { print "Average Size:", s/(NR-1)/1024, "MB,", NR-1, "Processes, Total usage:", (s/(NR-1)/1024)*(NR-1), "MB, Max Servers:", avail/(s/(NR-1)/1024) }'

Thanks for the idea and script. I modified it a little to accept the actual free resources from the output of free -tw

#!/bin/sh
available=$(free --mega -tw | grep "Total:" | awk '{print $4}')
ps -ylC httpd --sort:rss | awk -v avail="$available" '{ s += $8; } END { print "Average Size:", s/(NR-1)/1024, "MB,", NR-1, "Processes, Total usage:", (s/(NR-1)/1024)*(NR-1), "MB, Max Servers:", avail/(s/(NR-1)/1024) }'

Thank you for this! I have updated the original script with your suggestion.