netstat -t |grep 10050 |wc -l = more than 130 lines

[piotrlg]

Hi
I'm new zabbix user. Started from vmware appliance, and still testing it. I have installed agents on my Debian servers one 1.4 and one 1.8.2. On both servers when I do
netstat -t |grep 10050
.....
tcp 0 0 irys.rektorat.loc:10050 192.168.1.64:46256 TIME_WAIT
....

over 160 lines like above. Is it normal?

both servers are autodiscovered and connected to template_linux.
Well, on Zabbix 1.8 appliance server is the same situation. Sooo many sockets?
Regards
P.

[mattsmith]

Hi All,

I'd like to ask the same question as i have noticed the same issue.

It would be interesting to see how many connections others have.

I'm using:
Manually Added Host
server and agent 1.8.3
Linux_Template

Many Thanks
Matt

[dakol]

Hi
I'm new zabbix user. Started from vmware appliance, and still testing it. I have installed agents on my Debian servers one 1.4 and one 1.8.2. On both servers when I do
netstat -t |grep 10050
.....
tcp 0 0 irys.rektorat.loc:10050 192.168.1.64:46256 TIME_WAIT
....

over 160 lines like above. Is it normal?

a TCP connection in TIME_WAIT state is a dying connection.
It should go away after a timeout

Linux: sysctl net.ipv4.tcp_fin_timeout
Solaris: ndd -get /dev/tcp tcp_time_wait_interval

[mattsmith]

But why are there so many? Is one spawned for each item/check?

[dakol]

But why are there so many? Is one spawned for each item/check?

Zabbix Server established a tcp connection for each query (foreach item).

If you have 5 checks/second on this host, you will get 5*60 TCP in TIME_WAIT state

[mattsmith]

Cool, thanks for your help.