Service Console CPU Usage high

[TroyB]

How exactly do I troubleshoot this issue?

esxtop shows the Console at more than 90% used and top shows vmware-hostd being the culprit.

Now, I THINK this is related to the HP Agent but I'd like to learn how exactly I would find that out for sure.

As an example, is Windows I would take a look at the processes and could quickly tell EXACLTY which process is causing the spike. Anything more specific like that in Linux?

[doctormiru · Accepted solution]

Hi

3rd Party agents like HP, IBM or some security scanners can cause the service console to consume 100% CPU on the COS assigned CPU. The cause for this is not really clear for me but as far as I can say it has something to do with the scanning of the local web UI of ESX. Anonymous accesses to the Web UI can cause the hostd process to consume a lot of CPU. Don't know why at the moment. Restarting the service mgmg-vmware[/b] solves the problem for a while.

btw, top[/b] shows you more detailed proc info of the COS than esxtop.

Regards

Michael

[admin]

You also want to check that you have enough space on your Luns, do

\# vdf

Make sure that no Lun is more than 95% committed.

[kharbin]

From the console run:

top c

This will run top and show the highest consuming tasks first in the list.

Ken Harbin

www.esXpress.com

[TroyB]

Thanks for all the responses guys!

The most space used on a LUN is at 38% so that's ok.

top c shows this as the culprit:

/usr/lib/vmware/hostd/vmware-hostd /etc/vmware/hostd/config.xml -u

How do I restart this: mgmg-vmware service to see if that helps? And will this effect any VMs currently on the host?

Message was edited by:

TroyB

[admin]

as root

\# service mgmt-vmware restart

it shouldn't impact running VMs on a standard install

[TroyB]

Thanks guys! Restarting mgmg-vmware seems to have resolved the problem for now but I'll keep an eye out.