I have a dual VIO environment for each managed system. If I run 'viostat -sys 1 1' on VIO (id=1) and then run the same command on VIO (id=2), I get inconsistent sets (most disks are the same between the two outputs but there are a few missing in both outputs from the other). So for example, I might see disk570 in the output from VIO (id=1) but not in the output for VIO (id=2) or I may see disk30 in the output from VIO (id=2) but not in the output for VIO (id=1).
Also, if I attempt to run the 'lspv -size' on each VIO, I only get a subset of the disks for the 'viostat' command.
Is there a more reliable way to get a complete listing of all disk sizes and other metrics on my managed system (including vdisks)?
Pinned topic Need Reliable HMC/VIO Command To Display Metrics For All Disks on Managed System
D.R.Saravanan 06000139DJ1 PostACCEPTED ANSWER
Re: Need Reliable HMC/VIO Command To Display Metrics For All Disks on Managed System2014-08-07T05:42:59Z in response to JCantu
I would like to answer this question in 2 fold:
1. Running viostat with an interval of 1 second is too granular and there may be inconsistencies in output between 2 VIO Servers. One possible option is to try with a higher sampling interval like 5 or 10 seconds. The viostat also reports only disks that the VIO server can see, if one of the VIO server can't see the disk then it won't be reporting that.
2. From HMC v810, HMC provides REST APIs to provide performance statistics including Network & Storage IO. Kindly refer the following link to get the API documentation & details of metrics:
The HMC metrics limited to the Disk Adapters and also the sampling interval is fixed to 30s.
Kindly let me know if you are looking for further details.