ServiceGuard and Native SNMP Agents/SubAgents

[Karthik Narayanan_4]

Hi,
My question is as follows:

1: When the cluster node goes down the native SNMP agent does not send a link down trap nor does it send a cold start trap.

2: The trap that does come in has the following oid. 1.3.6.1.4.1.11.2.3.2.3 which is from HP Open View. ( I think)

Is this correct?

Thanks
karthik

[Ashwani Kashyap]

There is some SG patching issue here which HP is going to release shortly . But HP is going to release patches for ver 11.09 and 11.14 which are platform release and not for any other version which are feature release .

what is your SG version ? I am running 11.12 and had the same issues and thats what HP said . They also suggested me to upgrade to 11.14 and then apply the patch when it is released which should be by Sep end.

[Karthik Narayanan_4]

Hi,
Thanks for the response.

I would like to know the official link to what you have just said.
Would you please post the link to it?

Thanks
karthik

[Carsten Krege]

Karthik,

perhaps this document from the technical knowledge base is matching your request?

date: 6/5/02
document description: cmsnmpd not sending trap hpmcSGClusterDown
document id: 8606264135

Problem Text

CR# JAGae28464
problem
SServiceGuard 11.09 and HPUX 11.0

When cmsnmpd is used for sending traps to application, Traps are
sent for hpmcSGClusterUp, hpmcSGPkgUp, hpmcSGIPAddrUp, hpmcSGNodeUp
when cluster is started and then hpmcSGSvcDown, hpmcSGPkgHalt,
hpmcSGIPAddrDown , hpmcSGPkgDown and hpmcSGNodeHalted when the cluster
is halted. But there is no trap for hpmcSGClusterDown.

ACSL Support

06/05/2002

Fix Text

fix
The fix for this defect will include generating a cluster down trap and
updating the hpmcClusterState MIB variable to "down" when a node halts
or fails and it's the last node in the cluster. If the cluster status
can't be retrieved because the other nodes in the cluster are
unreachable the cluster down trap will still be sent, but the
hpmcClusterState MIB will be set to "unknown" instead of "down". This
may cause a cluster down trap to be sent even when the cluster is up if
the cluster has become partitioned and all nodes in the cluster aren't
able to communicate with each other because network connectivity has
been severed.

A site specific patch for 11.09 will go out around the end of August
2002. This fix will also be submitted to:

SG 11.09 PHSS_27158
SG 11.14 PHSS_27725
SG 11.13 PHSS_27722
SG 11.15 and later versions



Carsten