VMware Cloud Community
Stilez
Contributor
Contributor
Jump to solution

CLI commands to watch hypervisor process I/O activity and stats in detail

I'm in the situation of setting up a new vSphere server and want to watch it to check it directly via CLI to keep an eye on its maintenance activities for a few days, that it's not hanging, that hypervisor activity is running nicely etc. The web UI gives me graphs and task info gives me percentages, but both of those aren't really what I need - the graphs aren't specific as to what files/processes are doing what and the percentages often stall at a fixed value for many hours before jumping forward (eg when staging or remediating) so I can't be sure it's doing anything in the background. Also everything except CLI runs on a server/client basis on a different machine, whereas CLI runs on the relevant host directly, cutting out scope for confusion and issues.

On Windows I can use Resource Monitor or Task manager to watch individual processes, and their disk or network IO (latest bytes/sec delta for each process), and in resource monitor I can watch exactly which files are in use and what's being done with them, pretty much. On BSD and Linux, iostat and similar give much the same ability.

Is there an equivalent way to get this insight using CLI on ESXi 5.5 or 6.0, and if so, what is it? Or how do other people handle it if they want to check in detail what the server is doing during maintenance and that it isn't hanging or silently stalled due to IO/networking/SAN connectivity  issues?

0 Kudos
1 Solution

Accepted Solutions
TheBobkin
Champion
Champion
Jump to solution

Hello,

Likely what you want to be using here is esxtop run via SSH session to the hosts.

depping does a very nice summary guide of this tool and the options available:

http://www.yellow-bricks.com/esxtop/

If you wish to gather data over time using this you can set up a cron job using the steps outlined here:

https://kb.vmware.com/kb/1033346

You can also tell if the host is active or not "silenty stalled" by using tail command against /var/log/vmkernel.log or simply entering 'dmesg' from CLI.

Bob

-o- If you found this comment useful please click the 'Helpful' button and/or select as 'Answer' if you consider it so, please ask follow-up questions if you have any -o-

View solution in original post

2 Replies
TheBobkin
Champion
Champion
Jump to solution

Hello,

Likely what you want to be using here is esxtop run via SSH session to the hosts.

depping does a very nice summary guide of this tool and the options available:

http://www.yellow-bricks.com/esxtop/

If you wish to gather data over time using this you can set up a cron job using the steps outlined here:

https://kb.vmware.com/kb/1033346

You can also tell if the host is active or not "silenty stalled" by using tail command against /var/log/vmkernel.log or simply entering 'dmesg' from CLI.

Bob

-o- If you found this comment useful please click the 'Helpful' button and/or select as 'Answer' if you consider it so, please ask follow-up questions if you have any -o-

Stilez
Contributor
Contributor
Jump to solution

Thanks. I was using SSH, but the guide and especially the hint about visualesxtop (which I hadn't known of!) was exactly what I needed. Thank you!

0 Kudos