We are using the VMware Virtual Machine (SOAP) Sensor and have looked at the PRTG manual for this sensor. It states that the sensor shows read and write latency but that "which channels the sensor actually shows might depend on the monitored device and the sensor setup".
We have added several VMs from the same vcenter and ESXi host using the same probe. In each case the VMs are the same OS and vmx version. However, some of them present disk latency statistics and some do not. Is there any reason for this behaviour?
Daniel
Article Comments
Thank you for this. I set the Session Pool settings as you suggested then all the VMs collected data for disk latency. I set it back after about 5 hours and all but one stopped collecting the data again. The one that continued to work was different to the original one that worked, which is interesting perhaps.
I have set the Session Pool setting back to "Create a new session for each scan" but this creates increased load, I think. Do you have any other suggestion? Is this a bug?
Apr, 2016 - Permalink
Hello,
Thank you for your reply.
It seems that cached data related to VMware sensors should be deleted.
Default location is "C:\ProgramData\Paessler\PRTG Network Monitor\Sensordata (NonPersistent)\VMware Sessionpool".
Please delete this folder (it will be recreated by the next VMWare-SOAP-Sensor running).
Does that help?
Apr, 2016 - Permalink
I deleted the folder and allowed a couple more data points to be collected and then I changed the setting back to "Reuse session for multiple scans". Unfortunately this has not worked completely - three VMs are working, and they are all on the same host. Two further VMs do not work and they are on two different hosts. The VMs that do not work report zeros for disk latency at each poll.
Apr, 2016 - Permalink
Hello,
We need some more information from your PRTG. For analysis, please open the sensor settings and enable "Sensor Result" / "Write sensor result to disk". Please wait for the next sensor scan (or scan manually.) This should create a file located in "C:\ProgramData\Paessler\PRTG Network Monitor\Logs (Sensors)".
Please open a support ticket and send the results there (please refer to this KB article in the ticket).
Apr, 2016 - Permalink
Hello,
Please perform the following on the sensors where latency is missing:
- in the Settings of the affected device, under "Credentials for VMware/XenServer" -> "Session Pool", please choose "Create a new session for each scan" (instead of "Reuse session for multiple scans");
- keep that option for at least half an hour and then switch to "Reuse session for multiple scans".
This should refresh the cached data on the PRTG-machine.
Is there any change afterwards?
Apr, 2016 - Permalink