Hardware Status tab update frequency

[gzulauf]

How often does the data displayed on the Hardware Status tab of the vSphere Client update?

In the upper right-hand corner it indicates the last time it was updated and there is the option to manually force a refresh. But there does not seem to be a consistent frequency at which vCenter is polling for the hardware status data. On some of our servers it appears to be every few hours, others whole days can pass before a refresh.

Is there an option in the vCenter configuration to change the default polling frequency? We would like the data to be updated in parallel to our external monitoring systems (Nagios).

[DSTAVERT]

The HP tools are all we need for hardware monitoring.

Actually HP SIM relies on the same CIM modules that vCenter does.

[DSTAVERT]

SNMP is still an option which is supported by most monitoring tools.

[UniOfWyo]

DSTAVERT wrote:

SNMP is still an option which is supported by most monitoring tools.

Absolutely.   It would still be nice for things to work.  I think it's the job of the community to point out the things that don't

[DSTAVERT]

Not suggesting that you shouldn't point out things that don't work. Your post mentions not wanting to rely on one product for monitoring and since vCenter, HP SIM, Dell OpenManage and presumably what UCS uses all rely on CIM that SNMP was something that could be used in addition to CIM.

[wdugas]

All,

I opened a support request with VMware around the beginning of the year because the Hardware Status page of the VMware vSphere Client wouldn't automatically update the status, I finally got the following reply from VMware’s support today:

The Hardware status page acts solely as a CIM-client or consumer for the CIM-server on the ESX hosts. The vSphere client will retrieve the contents of this page from the Tomcat engine (on vCenter) which caches a copy of the page from an HTML/JSP engine. There are a timer coded inside this page that is responsible for updating the page at every 5 minute interval. There are a number of conditions that can cause the timer to stop, not limited to just those below:

1. When ESX host present in VC inventory was disconnected and reconnected back again.

2. When host is restarted, or powered-off and on again.

3. When sfcb within the host is stopped, in this case Hardware-status-tab doesn't get the latest data from host.

4. ESX host was removed from network and added again.

5. ESX host name resolution problems, and other problems related to network.

Since a condition exists which causes the timer to stop, the "updated" time on this page ceases to be updated until the timer is reset, the only valid method to reset the timer is to click on the "update" link.

Moving forward, in vSphere 5, the timer is intentionally reduced to a maxium lifetime of 1 hour (and not due to the timer being known to stop under certain conditions but for other reasons). The same steps will be required to update the CIM Client page when viewed through the Hardware Status tab in vCenter.

The good news is that this CIM client page is not tied into or related to the vCenter alarm mechanism in any way. Any issues with the alarm mechanism not triggering will be handled separately from the hardware status tab.

Oh, and don't forget at not all of the VMware alarms work, and VMware won't tell me which ones don't work...

Enjoy!

[srwsol]

This problem still exists in ESXi 5.  Also it looks like alarm conditions that go from green to red work without the page being updated, but when the alarm condition goes from red to green you get no notification unless the page gets updated first.  I tested this with the redundant power supply.  If the page isn't up or isn't autorefreshing and I pull the plug on one of the power supplies I get an email.  However, if I put the plug back in I don't get an email unless the page gets refreshed to reflect the changed status.  I tried pulling the plug, got the email, and then let the page quit updating automatically.  I then put the plug back in and left everything alone overnight.  No email.  Next morning when I manually refreshed the page, bingo instant email about an event that happend the previous night.  How lame.

Does anyone know if there is a CLI command that does what the clicking the update button does on the hardware page, either directed at the host itself or through vcenter?  I was looking for it and so far I haven't been able to find it.   If I do ever find it then I'll just write a script on a timer that does it every few minutes indefinately.

I can't for the life of me understand why they can't, or won't, fix this.  It's crazy.

Scott W.