Agent Ping no Data after Upgrade from Windows Agent 4.x to Agent2 6.x

[bernd.dausch]

Hi All,

we have Upated some (round about 20) Windows Agents from Version 4.x to Agent2 6.0. We use only Active Checks and Trapper.

Since the Update, two of the Server have the "Zabbix agent on Host not available for 5 Minutes". There are no Errors in the Agent logfiles,
all seams working fine. The other checks (Processer load (1 minute) have no problem and values every minute.

Only the Agent Ping has no values.

Some Ideas where to look and how to troubleshoot this?

Kind Regards,

Bernd

[Markku]

First is to check that the items are really properly configured and enabled.

Are the items configured using some template (= identically to all hosts) or have they been configured manually host by host?

If you say that ~20 hosts were updated but only two are failing and for this one item only, what could be the logical explanation for that? What is so special with those two hosts?

Are the other servers sending the agent.ping values properly every single time, or are they just sending them well enough not to cause the triggers to raise?

What's the agent 2 version, can there be bugs that have been resolved in later 6.0.x release?

Raising the logging level on the agent is a way to get more information about what it is doing. Traffic capture (tcpdump on the Zabbix server side and/or Wireshark on the Windows side) would give exact information, which checks did the agent receive from the server to execute, and what kind of results it produced. (Unless there is TLS involved, then it is more complicated.)

Markku

[bernd.dausch]

Hi Markku,

thank you for the quick response.

All Servers use the Same Templates and are configured identical.

On the other Server, I can't see missing values, It seams they send the values properly

We use the Version 6.0.9 of the Agent 2. I have found no bug report about this event.

We have raised the logging but found nothing suspicious for us. It looks like he is sometimes not sending the trigger without something in the log files.

I will collect the tcpdump and look if is something in there.

thank you

Bernd

[bernd.dausch]

I have attached a screenshot from the Agent Log. The Ping reported the Data each Minute then there is a pause from 9 minutes and then it is working again

[Markku]

It looks quite clear that the agent decides to not send the results for some time.

The seconds are also remarkable: it always sends 17 seconds past full minute, except during the problem time.

I don't have any explanation for this. Have you already reinstalled the agent? Is there any endpoint security software installed that maybe needs some adjustment?

Markku

[bernd.dausch]

Alle Server are on the same Virusscanner, but its a create Idea we will remove the antivirus and check again.

Thank you

Bernd

[bernd.dausch]

The Virusscan was not the issue, the problem still exist. I uninstalled the Agent2 6.0.9 and switched to 6.2.3, perhaps it is fixed in the current version.

Bernd

[bernd.dausch]

The Issue still exist with Version 6.2.3. Any Idea what to check next?

Bernd

[Markku]

Does increasing the agent logging even more reveal anything related?

How is your ServerActive configured, using IP or hostname?

Markku

[bernd.dausch]

Nothin special only infos which values were send, processing and sending of the values always without failures.

We have all Servers and Proxys configured with the IP Address

Bernd

[Markku]

Have you checked Windows event logs, are there anything suspicious during the time that the agent holds from sending the data?

Markku

[bernd.dausch]

There was nothing Special in the Windows Eventlog. The Server hat his Maintenance Windows for the Microsoft Update October last Weekend.
At the moment all is working as expected. I filter for the Agent Pings and monitoring it today. Perhabs, it is gone with the Windows Updates.

Bernd

[bernd.dausch]

Unfortunately the issue is back. Some other hints?

Bernd

[bernd.dausch]

We switched to the "classic" Agent Version 6.0.9 and there all is working fine. After switch back to the Agent 2, the problem comes again.

It seams this is a problem with the agent 2 version. What can we to next?

Kind Regards,

Bernd

[Markku]

You could open an issue in https://support.zabbix.com (unsure if they will handle it without support contract but worth trying, agent 2 penetration is maybe not huge yet so maybe they are interested in the case)

Markku