HP SIM reporting "A network interface has failed or redundancy has been lost"

[sphar]

Hi all,

HP SIM 6.3 reports this as a "Major Health Status" despite the fact that only one NIC is enabled for networking on each of our HP Proliants.

Has anyone found a way to rectify?

Thanks.

[BPE]

If you go to the system management home page you should see that the unused NIC are generating the error. orange color. if you ware running Windows you only need to disable the unused nic. But take care that you still have at least one working connection;-)
The health status does not change immediately. You need to wait 2 to 11 minute depending on your polling interval and the number of system in your SIM HW monitoring.

[sphar]

Thank you for replying.

I've disabled the unused NICs on the HP Proliant Servers but the health status / WBEM / SNMP reports several HP Proliant Servers "Management Processor Status" (iLo) as being 'degraded' or "SNMP passthru data".

Hope someone has observed this activity and has a resolution.

[infraune]

Hi all

how do same thing with esxi 4.1 host ?

thanks for reply

[shocko]

So its the iLo card that's generating the degraded status? In that case, if your iLo is disconnected then disable monitoring of it in the SMH (if using Wbem).

[infraune]

The problem is that i see 2 more network connections in SIM ( see attachment )

In VCM i put only 6

i dont know how to disable monitoring

[infraune]

same major status health with ESXi 5.0

[MariusGillesen]

I have only seen a workaround for "hiding" the unused NIC in ESXi4.1

1. Go to the ESXi host in VSphere

2. Select the Configuration tab

3. Go to Advanced Settings

4. Edit the settings and mark the unused NIC(s) for passthrough.

5. reboot the ESXi host.

This gets rid of the degraded messages and I have not seen any negative affect on our 30+ node ESXi cluster.

Not been able to test on ESXi 5, since we are a few weeks away from deploying that on HP blades.

[Bart_Heungens]

Hi,

Your workaround, was this for regular NICs or Virtual Connect?

My G6 10Gb NICs are connected to Virtual Connect modules where I have a profile for 2 virtual NICs assigned. The 6 other virtual NICs report now Not Connected and give this error in SIM.

However I cannot select these 6 NICs for passthrough. I can only do this on physical NIC which means that all 8 virtual NICs will be redirected... Which is not a solution in this case...

Kr,

Bart

[Rune Darrud]

Also note that not all hardware supports passthrough (DL385 G5).

Any news to this? I've seen this issue on ESX 3.5, ESXi 4.0, ESXi 4.1 U1, ESXi 5.0 (Current).

Just done upgrading to HPSIM 6.3, had hoped this had been resolved by now.

[SwisspostIT]

HP told us that another workaround (for VC environments) is to connect a dummy VLAN (internal only) to the unused NIC Ports but since this reduces the speed on the other ports this isn't a solution for us.

Fact is, that actually this is correct that HP SIM recognizes this as an failure since the system cannot differentiate between the NIC connectivity loss between a cable which isn't connected (and isn't needed) and a cable which is connected but defective...

In my opinion VMWare should add the option to ESXi to disable unused NICs like you can do this in Windows environments...

Regards,

Ville

[Rmitchell_1]

I just did a fresh install of SIM 7.0 and am just using it for the Insight Control for vCenter plugin.  This still gives the same error.  We only use 1 10G pipe for each adapter rather than carving the 4 Flex10/Flexfabric adapters.  I once put in a support call...over a year ago...and got a hold of someone from developement and told them why this annoyed us.  He said other customers compained and they were working on a fix...this was 6.2 days.  I'm sorry but this is HPs issue.  An unconnected network adapter should not be flagged as an issue.  I'm in the same boat as other users when it comes to Rack servers...we don't use ILO and therefore we always get issues when using SIM for those esxi hosts...which is why we don't use SIM other than for the vmware stuff on blades.  C'mon HP...fix this crap!!!

[Bart_Heungens]

Hi,

Some things I do not agree with, except the fact that it is annoying and should be solved...

This is not a HP only issue... In a Windows environment I can disable interfaces and the alerts go away... In ESXi I cannot disable interfaces... I think personally thatthis should be possible... If even Windows can do it?

Further I do not agree with your statement "An unconnected network adapter should not be flagged as an issue". Why do you want to do monitoring then?

For me it is really important that know that there is a network problem ,since it is another team that manages it...

Had several times that my servers were running on 1 NIC and teaming took over, I prefer however that both NICs are active...

Kr,

Bart

[Patrick Viquerat]

Hi,

If you are using HP SIM v 7.0 to monitor your ESXi servers, you can disable the status of unconnected NICs:

- select the ESX

- go to Tools, --> Configuration Tool, --> ESXi Component Status

- select Run Now

- select the NICs you want to ignore the status and click on "Ignore Status"

Wait a few minutes and your server will be green again :-)

Patrick

[Bart_Heungens]

Hi,

This solution makes indeed the server green again, but it is not ideal yet.

The server still gets orange via the ILO. The ILO still reports (on my Gen8 blade for sure - didnt test it yet on G6/7) an orange state saying that 1 or more NIC interfaces are down...

The search continues...

Kr,

Bart

[Simkill]

Do you really need SNMP pass thru on the iLO interface? I turned all mine off to avoid this annoying message by logging into the iLO and turning it off under administration.

[steven Prinz]

It would be extremely helpful if HP could put functionality into SIM wherein the status of selected NIC(s) could be ignored. I'm not a programmer, but it seems like that would be a fairly easy feature to add. This seems like the more reasonable approach, since the solution would be applied to the hardware being monitored, no matter what operating system is installed. Making a change to the configuration of the installed operating system should not be required in order to monitor the hardware status. To me, disabling the unused NIC(s) in Windows is a workaround to the problem, not a solution.

[Martin2000]

In SIM 7 I was able to disable monitor status of individual nics by:

In SIM console select ESX host -> "Tools & Links" tab -> "WBEM Health Inclusion Status" link -> select 1 or more nics and select "Ignore Status"

You will see Propagation State change to false.

[drpixel]

Hello,

The Major status for the iLO is often related to some major events in the IML of the server.

Just clear it (or mark them as repaired) and the status will come green again.

To verify it quickly, just perform an hardware status polling on the server (after you've "repaired" the IML)

[zeroagemain]

I've been having this issue too and Martin2000's advice worked for me after an IC upgrade to 7.1, thanks.

[jloberle]

Does anyone know if there is a way to script this. We have over a 100 host and this will take forever to do manually.

[TODD SMITH]

You can get Insight to ignore ESX disable devices, its easy,

In Insight Select the ESX server, Under tools & Links, Select "WBEM Health Inclusion Status"

On the "WBEM Health Inclusion Status" screen select the devices that you want to ignore then click ignore status.

You may not see a change for up to 11 mins or so. Works Great but you have to do this on each ESX Server.

Just remember is you so the ignore devices, come back and select and click the include status. 

Best of Luck

[EMEASupport]

We simply add loopback adapters to any unused NICs - would this not be a solution?

[rob_m]

Does anyone have a solution with this issue on Linux servers?  I don't have the loopback plugs.

[woby]

---

@MariusGillesen wrote:

I have only seen a workaround for "hiding" the unused NIC in ESXi4.1

1. Go to the ESXi host in VSphere

2. Select the Configuration tab

3. Go to Advanced Settings

4. Edit the settings and mark the unused NIC(s) for passthrough.

5. reboot the ESXi host.

This gets rid of the degraded messages and I have not seen any negative affect on our 30+ node ESXi cluster.

Not been able to test on ESXi 5, since we are a few weeks away from deploying that on HP blades.

---

very helpfull answer, safed me some time, thank you :)