metrocluster error: Failed to get local DR group state from the SMI-S (CIM_ERR_FAILED)

[ARETESOP]

Hello, I Am mounting metrocluster on 2 sites (with EVA4400, 2 blades with HPUX 11.31 and MCSG 11.19) each one. The packages are IVM 4.1. I am still the guide B7660-90026 (march 2009,Cap4 Building Disaster Tolerant Serviceguard Solutions Using Metrocluster with Continuous Access EVA.
DRG replicated and ok.
Smiseva.conf edited and ok
smispasswd ok
mceva.conf edited and ok
evadiscovery-f ok
evadiscovery-l ok (generated map is correct).
caeva.env distributed to pkg_dir on all nodes.
Cmmakepkg - m dts/mccaeva edited to include the package generated by IVM's toolkit.
cmrunpkg start and stop de pakage, the log of the package shows the following mistake:

Nov 17 21:14:47 root@bl1r6 master_control_script.sh[5721]: ###### Starting packa
ge pkgr6 ######
Nov 17 21:14:47 root@bl1r6 mc.sh[5764]: Starting Metrocluster
Nov 17 21:14:47 root@bl1r6 caeva.sh[5772]: Starting Metrocluster CAEVA Package
Nov 17 21:14:47 - Node bl1r6 : Entering Data Replication Enabler
Nov 17 21:14:47 - Node bl1r6 : Connecting to SMI-S server 10.2.0.14:5989 using H
TTPS connection.
Random Number Generator: /dev/random
Nov 17 21:19:46 - Node bl1r6 : Error - Failed to get local DR group state from t
he SMI-S 10.2.0.14;
CIM_ERR_FAILED
Nov 17 21:19:46 - Node bl1r6 : Error - Unable getting the Data Replication state
or the state
is not valid. Fix the problem and then restart t
he package
Nov 17 21:22:05 - Node bl1r6 : Leaving Data Replication Enabler
Nov 17 21:22:05 root@bl1r6 master_control_script.sh[5721]: ##### Failed to start
package pkgr6, rollback steps #####
Nov 17 21:22:05 root@bl1r6 caeva.sh[6310]: Stopping Metrocluster CAEVA Package
Nov 17 21:22:05 root@bl1r6 mc.sh[6318]: Stopping Metrocluster

Any idea over this error?
The SMS EVA sees 2 perfectly, and the evadiscovery used the port 5989 to connect and to generate the map.

Thanks in advance

[Stephen Doud]

Make CERTAIN your cluster meets compatibility requirements identified in this document:

Serviceguard Disaster Recovery Products Compatibility and Feature Matrix (Metrocluster Continuous Access EVA), September 2009 HP-UX 11i v1, 11i v2, 11i v3 [ PDF: http://docs.hp.com/en/6192/DT-matrix-mceva.pdf ]

An incompatibility between SymCLI and Metrocluster/EVA can produce strange errors.

Also insure the SymCLI database contains the same usernames identified in smiseva.conf

[ARETESOP]

The prerequesitos are correct
Metrocluster A.04.00 mar 09
HPUX 11iv3 mar 09
WBEM A.02.07.04
Evainfo 9.0
MCSG A.11.19
EVA4400 XCS09522000
SMI-S EVA v9.0
CV 9.00.01
Virtual Machines 4.1

attched file txt with more info

[Stephen Doud]

Metrocluster uses the evadiscovery command to talk to SMIS to get array status.

On page 167 in the Sep 2009 edition of the Desiging Disaster Recovery HA Clusters document at http://docs.hp.com/en/B7660-90027/B7660-90027.pdf
it states:

NOTE: Before running the evadiscovery command, the management server configuration
must be completed using the smispasswd. Otherwise, the evadiscovery command, will fail.
NOTE: Run the discovery tool after all storage DR Groups are configured or when there is any change to the storage device. For example, the user removes and recreates a DR group that is used by an application package. In this case the DR Group's internal IDs are regenerated by the
EVA system. Update the external configuration file if any name of storage systems or DR groups is changed, run the evadiscovery utility, and redistribute the map file /etc/dtsconf/caeva.map to all Metrocluster clustered nodes.

If 'evadiscovery -l' fails, start on page 155 and review your Metrocluster <-> EVA communication preparations.