Monitor connection between vROps and vCenter servers

[wbabineaux]

Hi all,

After an issue we had a few weeks ago, VMware noticed this from the logs:

2018-01-24 10:37:42,038 ERROR [VMVI RetrieverThread Datacenter Watcher:1]  (82850) com.vmware.vcops.vlsi.VcSessions.terminateSession - Unable to logout from VC

(vim.fault.NotAuthenticated) {

   faultCause = null,

   faultMessage = null,

   object = ManagedObjectReference: type = Folder, value = group-d1, serverGuid = null,

   privilegeId = System.View

.....

And further down, it looks like it tried to reconnect but could not:

2018-01-24 10:37:42,038 WARN  [VMVI RetrieverThread Datacenter Watcher:1]  com.vmware.vim.vmomi.client.http.impl.HttpConfigurationCompilerBase$ConnectionMonitorThreadBase.shutdown - Shutting down the connection monitor.

2018-01-24 10:37:42,038 INFO  [VMVI RetrieverThread HostMemorySystem:3]  (82850) com.vmware.vcops.vlsi.VcSessions.terminateSession - Terminating the VC connection.

2018-01-24 10:37:42,038 WARN  [VLSI-client-connection-monitor-84]  com.vmware.vim.vmomi.client.http.impl.HttpConfigurationCompilerBase$ConnectionMonitorThreadBase.run - Interrupted, no more connection pool cleanups will be performed.

2018-01-24 10:37:42,038 INFO  [VMVI RetrieverThread Datacenter Watcher:1]  (82850) com.integrien.adapter.vmware.VMwareAdapter.connect - Creating connection Datacenter Watcher:1, with socket timeout:3600000

2018-01-24 10:37:42,038 INFO  [VMVI RetrieverThread HostMemorySystem:3]  (82850) com.vmware.vcops.vlsi.VcSessions.terminateSession - Attempting to get the session

2018-01-24 10:37:42,047 ERROR [VMVI RetrieverThread HostMemorySystem:3]  (82850) com.vmware.vcops.vlsi.VcSessions.terminateSession - Unable to logout from VC

(vim.fault.NotAuthenticated) {

   faultCause = null,

   faultMessage = null,

   object = ManagedObjectReference: type = Folder, value = group-d1, serverGuid = null,

   privilegeId = System.View

I was informed this was a disconnect between vROps and the vCenter server.

Question:

I wanted to ask if there was any type of monitoring that I could create when vROps loses connection to a vCenter server?

How can I monitor for the disconnect?

     - By a specific alert

     - By a port disconnect

     - By a ping/SNMP trap?

Thanks for any and all help!

[daphnissov]

Question:

I wanted to ask if there was any type of monitoring that I could create when vROps loses connection to a vCenter server?

How can I monitor for the disconnect?

     - By a specific alert

     - By a port disconnect

     - By a ping/SNMP trap?

Hmm, this is odd. But I'm betting you could monitor this by using vRLI to watch for that event. You'll need to configure the liagent on the vROps node(s) directly to point to vRLI, but these are pre-installed for you. You probably want to update them, however, but this is how I'd tackle it.

[wbabineaux]

I spoke with a few individuals on the Virt team and they tried using the version of vRLI that came with our licensing. I don't believe this was going to fit our use case. As of right now we do not have vRLI and do not plan on getting it anytime soon.

As far as the connection between vROps and vCenter, how do they connect? How do they talk?

I know the VM agent is used to pull the metrics, but what port is vROps using to talk to vCenter over? I assume I can monitor for this connection going down between servers?

I took a look here, but did not find the correct information:

https://pubs.vmware.com/vrealizeoperationsmanager-62/topic/com.vmware.ICbase/PDF/vrealize-operations...

[daphnissov]

As far as the connection between vROps and vCenter, how do they connect? How do they talk?

It's just over the API. The integration built into vRLI allows you to supply credentials and have the solution pull tasks and events directly from vCenter. For added visibility, you can also configure syslog on the vCSA to send its syslog streams over to vRLI. But for even more flexibility, you can install the vRLI agent on the vCSA and send over any other log streams you may need to capture. Doing so uses the native CFAPI port to vRLI over TCP 9000.

[sxnxr]

This connection is monitored by default.

When you log into vrops you will see the alarm bell at the top right. This is for all your adapters ( the below is our lab hence they are alerting) did you have any alerts in here when you had the problem.

You can create your own alert to go where you want by using the below symptoms

This may have alerted when you were having problems

[wbabineaux]

Sorry for the late reply, as I have been busy with new build outs and work with other solutions.

We did not have alert for this issue within vROps. I also have the adapter down alert set to send an email, and I did not receive an email either.

The only thing that showed a disconnect between vROps and the vCenter server was the error in the log file.

I have went ahead and set up outbound notifications for those specific symptoms to go along with the Alert emails for an adapter instance being down.

I am in the process of trying to get vRLI for use to use and possibly monitor for this issue as well as others.

Thanks for all the answers help!