Fictitious loss of ILO "network link"

[Derek Balling]

Since upgrading to hp-health 8.7.0, we've been seeing  lot of:

Trap-ID=9015

The Remote Insight/ Integrated Lights-Out firmware has detected the loss of network link.

 It gets immediately followed, within a minute or two with:

Trap-ID=9016

The Remote Insight/ Integrated Lights-Out firmware has detected the presence of network link.

 We originally thought it might be a firmware issue, but it doesn't seem to be tied to any particular firmware rev.

Anyone seen this, or have any thoughts?

[Derek Balling]

Bumping this to see if anyone else is seeing this? We're seeing it across the board on all of our servers that got the latest hp-health and snmp-agent RPMs installed on them.

[Papasmurf]

I've been seeing a lot of these as well.  They do indeed appear to be ficticious and do not corespond with any measurable network outage as the active interfaces remain pingable.  I wonder if it has anything to do with unconfigured interfaces.

[Derek Balling]

I don't think it has to do with unconfigured interfaces because these are specifically the ILO NIC, at least according to the MIB:

   cpqSm2NicLinkDown TRAP-TYPE
        ENTERPRISE compaq
        VARIABLES  { sysName }
        DESCRIPTION
            "The iLO NIC Link is Down.

            The Remote Insight/ Integrated Lights-Out firmware has 
detected the loss of network link."

              --#TYPE "The iLO NIC Link is Down (9015)"
              --#SUMMARY "The iLO NIC link has been lost."
              --#ARGUMENTS {}
              --#SEVERITY CRITICAL
              --#TIMEINDEX 99
              --#STATE   DEGRADED

        ::= 9015

And I see this on blades, where there's absolutely no way the ILO NIC (which is fed on the backplane) is seeing a link drop condition. One that ONLY a single blade at a time is seeing.

[chinmay]

Let me check on this, will revert soon

[MARREEL Chris_1]

I can confirm we have the same issue.

1 Blade Chasis.  7 ESX-hosts and 1 W2K8R2 host.

And HP Sim is reporting this problem allso.  But there is no real problem with the ILO-boards, so this is a fake error message.

What can we do to solve this ?  Anyone found a sollution ?

Greetings,

  Chris MARREEL

[MARREEL Chris_1]

I have investigated this problem at our site a bit further.  The firmware was up-to-date, but the HP-agents that were loaded on the ESX-hosts were still version 8.70.

and I have updated them to the 9.10 (current version), and since this update the error about the ILO connectivity is gone.

So I hope this can help others to solve this problem if they still encounter the problem.

Greetings,

  Chris MARREEL

[JPansanel]

Hi,

We have exactly the same error with G6 blades. It does not occur with G7 (same RPM version).

We have updated iLO version and RPM, but the problem remains.

Cheers,

Jerome Pansanel