zabbix-proxy snmp communiction error

[ngabor]

Hi!

I have a problem.

I use zabbix/zabbix-server-mysql:alpine-3.4.10 docker image and zabbix/zabbix-proxy-mysql:alpine-3.4.10 docker image
The proxy is communicate with the server. It's OK.

When I settting the proxy on host configurations, then the SNMP "interface" not working.
The SNMP communication not working on the zabbix-proxy:

164:20180612:075907.646 received configuration data from server at "zabbix-server", datalen 235989
176:20180612:080023.768 SNMP agent item "areca.battery.level" on host "areca" failed: first network error, wait for 15 seconds
164:20180612:080908.394 received configuration data from server at "zabbix-server", datalen 235989
181:20180612:081024.022 temporarily disabling SNMP agent checks on host "areca": host unavailable
164:20180612:081908.582 received configuration data from server at "zabbix-server", datalen 235989

If I settings this host use the zabbix-server, the communictaion is working good.
This bad operatin is not available on all devices with snmp-interface (FREENAS).

thy.
G

[kloczek]

176:20180612:080023.768 SNMP agent item "areca.battery.level" on host "areca" failed: first network error, wait for 15 seconds

I think that someone should change a bit this error message because it suggest that it may be related to some network layer issues ("network error").
This issue has been many times discussed on zabbix forums and many other places where people are discussing use SNMP.
This issue is related to some bugs on SNMP agent side (bugs in net-snmp code).

[ngabor]

I think that someone should change a bit this error message because it suggest that it may be related to some network layer issues ("network error").
This issue has been many times discussed on zabbix forums and many other places where people are discussing use SNMP.
This issue is related to some bugs on SNMP agent side (bugs in net-snmp code).

yes, I reading .
But, this configuration on the zabbix-server working good with snmp-interface.
Only the zabbix-proxy not working snmp-interface (zabbix-server send configuration).

If I setup the all snmp-host (73 switch and 32AP) on the zabbix-server, then the server load is >8.
I would like to use proxy servers for the resources distribution, but is not working with some...

Bugs on SNMP agent only the zabbix-proxy?

[LenR]

So no host snmp works from the proxy? Have you tried the command line tools from the proxy to a failing host as a test? Can you ping the failing hosts?

"Network errors" include firewall blockages :-)

[kloczek]

Bugs on SNMP agent only the zabbix-proxy?

Possibility of something like this is very low because zabbix server and proxy are sharing exactly the same code on interacting with SNMP agents.

Try to switch temporary back to the server to confirm that on using exactly the same template or templates and he same number of StartPollers you are observing different effect.
And/or at least try to confirm that using snmp{,bulk}walk commands you are able to query the same set of oids from locations where are srv and prx.

[ngabor]

So no host snmp works from the proxy? Have you tried the command line tools from the proxy to a failing host as a test? Can you ping the failing hosts?

"Network errors" include firewall blockages :-)

no, some devices work, for example FreeNAS.
The snmpwalk command is working on the zabbix-proxy with the problematic host.

...

[LenR]

What is the timeout value in the zabbix proxy config vs. the timeout value on the zabbix server?

[ngabor]

Possibility of something like this is very low because zabbix server and proxy are sharing exactly the same code on interacting with SNMP agents.
Try to switch temporary back to the server to confirm that on using exactly the same template or templates and he same number of StartPollers you are observing different effect.
And/or at least try to confirm that using snmp{,bulk}walk commands you are able to query the same set of oids from locations where are srv and prx.

I enlightened

I did not think about the following:
This is my a docker swarm setup problem:

the docker swarm virtual subnet cime 10.0.0.0/16 and a
the real address of devices to be monitored 10.0.0.0/24
which is so obvious that it causes uncertainty.

I was delighted to have helped you in the awakening of thought. Always learns about everythin

I use the snmpwalk in the "for ..." cycle, after
the network problem was revealed.

Thanks for your time.
G
...