Web-server – Apache Connection timeout, what could be wrong

apache-2.4web-server

I have a Apache server which is very unstable, and many times a day I struggle to connect to apache. I can not figure out why, since it for me everything seems to be ok. The server have at all time "plenty" of free resources. The server is a Linode 4098 with 4GB ram, 48GB HD, and 2 CPUs.

I'm using mod_status to check apache status but I keep getting Connection timeout errors when i try access /server-status page.

My current apache conf

<IfModule mpm_prefork_module>
        StartServers            3
        ServerLimit             200
        MinSpareServers         3
        MaxSpareServers         5
        MaxRequestWorkers       175
        MaxConnectionsPerChild  100
        MaxRequestsPerChild     20
</IfModule>

Almost every second apache request to /server-status fails with ERR_CONNECTION_TIMED_OUT.

What I cant understand is why this happens, as when I check server there is plenty of available resources and idle connections.

For example /server-status reports many available connections:

Parent Server Config. Generation: 1
Parent Server MPM Generation: 0
Server uptime: 1 hour 6 minutes 6 seconds
Server load: 0.09 0.15 0.14
Total accesses: 27206 - Total Traffic: 79.8 MB
CPU Usage: u3.92 s.77 cu0 cs0 - .118% CPU load
6.86 requests/sec - 20.6 kB/second - 3077 B/request
52 requests currently being processed, 1 idle workers

RRRCRRRRRRRRRRRRRRRRRRCRRC.RR.R.RRRRRRRRRCRRRRR_WCCRR.R..RR.....
................................................................
...............................................

And with uptime, i see that CPU is also available

load average: 0.05, 0.12, 0.14

and there is available memory:

        total   used   free  shared  buff/cache   available  
> Mem:   3.9G    1.3G   413M     81M        2.1G        2.3G

Anyone have some idea about what could be wrong?

UPDATE:
As requested, here is a snippet of the access log. (99% is requests regarding tracking stats by calling /api2/counting/count script-handler)

4.139.128.194 - - [23/Jan/2017:11:38:36 +0000] "POST /api2/counter/count?r=1485171514298 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
176.38.149.76 - - [23/Jan/2017:11:38:36 +0000] "POST /api2/counter/count?r=1485171517077 HTTP/1.1" 200 454 "-" "WriteDiary4.72"
67.84.227.131 - - [23/Jan/2017:11:38:36 +0000] "GET /keepalive?471:1485171524866 HTTP/1.1" 200 469 "https://www.writediary.com/notes" "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/51.0.2704.79 Safari/537.36 Edge/14.14393"
122.178.156.252 - - [23/Jan/2017:11:38:36 +0000] "POST /api2/counter/count?r=1485171515804 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
82.137.13.18 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171525558 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
103.199.35.221 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515470 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
95.218.142.245 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171516264 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
47.31.143.106 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515787 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
94.139.128.194 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515619 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
120.188.33.198 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515867 HTTP/1.1" 200 454 "-" "WriteDiary4.72"
47.29.174.163 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515122 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
85.115.224.151 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515453 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
47.11.219.147 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171516596 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
61.5.53.52 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171520665 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
116.102.205.20 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171516590 HTTP/1.1" 200 454 "-" "WriteDiary4.72"
101.60.31.133 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171516802 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
47.31.147.213 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171515431 HTTP/1.1" 200 5131 "-" "WriteDiary4.72"
186.78.73.222 - - [23/Jan/2017:11:38:37 +0000] "POST /api2/counter/count?r=1485171669305 HTTP/1.1" 200 454 "-" "WriteDiary4.72"
112.79.186.36 - - [23/Jan/2017:11:38:38 +0000] "POST /api2/counter/count?r=1485171736857 HTTP/1.1" 200 454 "-" "WriteDiary4.72"
94.139.128.194 - - [23/Jan/2017:11:38:38 +0000] "POST /api2/counter/count?r=1485171516605 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
24.236.88.161 - - [23/Jan/2017:11:38:38 +0000] "POST /api2/counter/count?r=1485171516752 HTTP/1.1" 200 441 "-" "WriteDiary4.72"
150.31.100.192 - - [23/Jan/2017:11:38:38 +0000] "POST /api2/counter/count?r=1485171517315 HTTP/1.1" 200 454 "-" "WriteDiary4.72"
12.181.196.67 - - [23/Jan/2017:11:38:38 +0000] "GET /keepalive?50412:1485171517975 HTTP/1.1" 200 501 "https://www.writediary.com/notes" "Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/54.0.2840.99 Safari/537.36"
112.198.75.139 - - [23/Jan/2017:11:38:38 +0000] "POST /api2/counter/count?r=1485171408169 HTTP/1.1" 200 4794 "-" "WriteDiary4.72"
114.125.201.114 - - [23/Jan/2017:11:38:38 +0000] "POST /api2/counter/count?r=1485172944904 HTTP/1.1" 200 441 "-" "WriteDiary4.72"

SOLUTION:
I had to upgrade server to allow more workers.
As each Apache connection reqires about 30MB or Ram, the server was pushing the limits 4GB RAM gives = 4000MB / 30MB = 133 connections max… and when some of the memory is allocated to mysql and other system functions, in practice only 3.5GB is available, giving a max closer to 100 connection.

When upgrading to 8GB RAM, max limit is closer to 250 connections, which makes everything run smooth. In fact, after rescale RAM, server-status reported 200 requests was simultanously processed by Apache for many hours in peak time. Thats would explain why the server struggeled in peak time with a limit closer to 100..

Best Answer

can you please post the apache access log

you can find logs if you are using ubuntu by using this command

cat /var/log/apache2/access.log

but it varies according to your platform

Related Topic