We are running zabbix 2.0.6 on Ubuntu 12.04 x86_64.
We have about 350 hosts with 2500 items total.
Server is remotely located and all remote networks are connected via vpns, have latency about 100-150 ms.
We have some simple checks on linux hosts, and many SNMP items from different hardware.
Strange thing is that after some days server stops receiving any udp packets. It can't connect to snmp, even DNS become broken, can't resolve any host.
Only reboot helps.
At this time we can see udp errors in netstat:
This sample just after rebooting, we have zero errors.
We tried different advices, tuned buffers etc, tried another kernel and even another server, the problem persists.
This problem is definitely caused by Zabbix, but we can't solve it for several months.
We have about 350 hosts with 2500 items total.
Server is remotely located and all remote networks are connected via vpns, have latency about 100-150 ms.
We have some simple checks on linux hosts, and many SNMP items from different hardware.
Strange thing is that after some days server stops receiving any udp packets. It can't connect to snmp, even DNS become broken, can't resolve any host.
Only reboot helps.
At this time we can see udp errors in netstat:
Code:
root@freedom1-hetzner ~ # netstat -su
IcmpMsg:
InType0: 16954
InType3: 906
InType8: 37
InType11: 1
OutType0: 37
OutType3: 149
OutType8: 19568
Udp:
40094 packets received
162 packets to unknown port received.
[B]0 packet receive errors[/B]
16838 packets sent
We tried different advices, tuned buffers etc, tried another kernel and even another server, the problem persists.
This problem is definitely caused by Zabbix, but we can't solve it for several months.