cmcheckconf gives protocol error

[Natesan_1]

While running cmcheckconf we are getting following error,

Checking clluster file: /etc/cmcluster/cmclconfig.ascii
Checking nodes... Done
Checking existing configuration...
Done
Gathering existing configuration information...........Done

Protocol failure with cmclconfd on au2 (1) : I/O error
Error: unable to determine network configuration:failed to setup for probing the network on node au2: I/O error
Failed to probe network
cmcheckconf : unable to reconcile concifugration file /etc/cmcluster/cmclconf.asii with discovered configuration information.

we have checked,
1. linkloop between & across port
2. lanscan for any duplicate MAC address
3. we tried run cmquerycl one lan port (lan5 FDDI) not coming in the output also it was throughing error like unable to reach port
4. we tried cmcheckconf removing that particular port we are not getting above error.

asper ITRC doc: A5100810, it has been told that duplicate MAC address.

Now customer disagree for the duplicate MAC address.
he wanted congreet proof.

I was tried changing MAC of suspected lan and try but needs cmcluster to be brought down.

So any Gentle folk as more suggestion and diagnostics help etc or welcome.

Customer using MC/SG version:11.09, swlist attached for reference

Thanks in advance.

[melvyn burnard]

I would recommend you at least update the SG patch level to PHSS_27158, if not look at upgrading to a supported Sg version, such as 11.14 or 11.15

Also, run a cmscancl and read the contents of the output file to verify what Serviceguard reports.
You do have some form of networking issue that you need to fix, have you checked broadcast addresses?

[Sridhar Bhaskarla]

Hi,

Do you see any errors in syslog.log files of both the servers?.

Was the cluster already running or you are trying to create a new cluster?.

If you are not able to reach the FDDI ports, then it shouldn't be in your configuration file.

I would check the hostnames and their resolution on both the nodes. Try "rlogin" to the same host and to the other host on both the servers. Make sure none of the interfaces are with 0.0.0.0 IP addresses. If you find them, then unplumb those interfaces.

-Sri

[Natesan_1]

Hi Melvyn,

Thanks for ur reply.

At present cluster is up and running customer was able to do changes and apply the some time before. Presently this error reporting...
1.Do you think it could be patch issues?.
2.I have checked all the stationary ip brodcast address, looks fine.
3.I have checked nettle log afte formating, could not find any error partaining network ports.

Any more suggestions if u have?,.