topic Error when launching cmcheckconf in Operating System - HP-UX

Error when launching cmcheckconf

Horia Barbul_2 — Tue, 02 Nov 2010 12:03:36 GMT

Hello.

I am trying to make a cluster out of two nodes made of make_sys_image of an original SG cluster.

When issuing the cmcheckconf command the following error appears:

# cmcheckconf -v -C /etc/cmcluster/cluster_drc.ascii
Unable to receive reply from local cmclconfd. Connection timed out
Unable to initialize HOSTNAME_ADDRESS_FAMILY.

The cmclconfd process was apparently properly started:

# ps -ef | grep cmcl
root 3299 2084 0 21:44:25 console 0:00 grep cmcl
root 2356 1061 0 20:56:07 ? 0:00 cmclconfd -c
root 2358 1061 0 20:56:07 ? 0:00 cmclconfd -c
root 2357 1061 0 20:56:07 ? 0:00 cmclconfd -c

Since these systems are in a test environment without DNS server all relevant name info are in /etc/hosts:

#
172.16.1.69 IgniteDR ignitedr ignitedr.cne.ro
172.16.1.66 iasdb1drc iasdbidrc.cne.ro
172.16.1.67 iasdb2drc iasdb2drc.cne.ro
172.16.1.144 iasdb1 iasdb1.cne.ro
172.16.1.145 iasdb2 iasdb2.cne.ro
172.16.1.147 cldb.cne.ro cldb#
172.16.1.148 clas.cne.ro clas
127.0.0.1 localhost loopback
and /etc/nsswitch.conf is modified to go first to "file".

The last lines in /var/adm/syslog.syslog.log are in the attached file

Any hint would be highly appreciated.
Thank you

Re: Error when launching cmcheckconf

Jayakrishnan G Naik — Tue, 02 Nov 2010 12:37:25 GMT

Hi

The attachment is not added here. So please do it. Have you added the necessary entries in cmclnodelist file?

Regards
Jayakrishnan G Naik

Re: Error when launching cmcheckconf

Horia Barbul_2 — Tue, 02 Nov 2010 12:46:25 GMT

Sorry for missing to attach the log file.

Re: Error when launching cmcheckconf

Horia Barbul_2 — Tue, 02 Nov 2010 13:08:04 GMT

Hi Jayakrishnan.

Since the nodes are made from the images of the nodes of a working cluster all needed information should be allready there:

# hostname
iasdb1
#
# cat /etc/cmcluster/cmclnodelist
iasdb1.cne.ro
iasdb2.cne.ro
cldb.cne.ro
clas.cne.ro
#

# hostname
iasdb2
#
# cat /etc/cmcluster/cmclnodelist
iasdb1.cne.ro
iasdb2.cne.ro
cldb.cne.ro
clas.cne.ro
#

Thank you
Regards
Horia

Re: Error when launching cmcheckconf

Matti_Kurkela — Tue, 02 Nov 2010 13:18:53 GMT

> Nov 2 21:01:27 iasdb1 cmwbemd[1622]: CIM_ERR_NOT_FOUND: HP_ServiceguardIndication

> Nov 2 21:06:55 iasdb1 syslog: icapd: Caught exception: ERROR: The following low-level WBEM error occurred: CIM_ERR_NOT_SUPPORTED: @31:The system is not using a compatible version of IPMI. The nPartition Provider supports IPMI 1.5.[The system () is using IPMI 1.0]

The error messages suggest the CIM/WBEM service is having some problems. I understand this is an essential component for nPar-capable systems.

> I am trying to make a cluster out of two nodes made of make_sys_image of an original SG cluster.

What kind of hardware was the original server using? Have these two nodes been restored to same model hardware?

If the hardware models are the same as in the original cluster, you might want to check the firmware versions. Perhaps the original cluster's patchlevel requires newer firmware levels than your current hardware has?

MK

Re: Error when launching cmcheckconf

Horia Barbul_2 — Tue, 02 Nov 2010 13:29:31 GMT

Hi Matti.

The "source" and the "destination machines are not identical. This is actually why I receive the CIM errors in syslog. It seems that NParProvider ( which was installed on the original rx7640 ) does not properly work on the destination servers which are not cell-based (BL870c) hence I get the CIM errors.

I tried

# cimprovider -r -m HP_NParProviderModule
PGC00614: Deleting the provider module...
PGC00607: The provider is deleted successfully.
#
but I still get the error message.

However I did the same tests before and I was able to verify and start the cluster ( with packages and all). This time I get the cmcheckconf error...

Thank you
Horia