SQL6048N - Error in Node B after Cluster FailOver

[Former Member]

Hi,

We have an SAP ECC 6.0 System running on AIX 6.1 and DB2 9.1 - HACMP cluster. We have installed the CI and DB on Node A in the below mentioned procedure.

1. Installation of ABAP Central Service Instance (ASCS00)

2. Installation of Central Service Instance (SCS01)

3. Installation of DB2 Software on /opt/IBM/ - Manual Installation of Db2 Software using db2setup in Database RDBMS Media.

4. Installation of Database Instance

5. Installation of Central Instance (DEVBMGS02)

All datafiles and SAP Files are on the Shared Disk in SAN, we have NFS mounted the /usr/sap/trans & /sapmnt/SID filesystems from the /export directory. The NFS mounted filesystems are available on both nodes.

We have installed the HACMP cluster after finding that the SAP System is starting up at Node A, with the Database and SAP Instance started up separately at Node A.

We have also copied all the Profile files, user environment files from Node A to Node B and done necessary changes to the file names such that the SAP Instance in Node B after failover is able to use the files.

After shifting the Resource Groups to Node A, when we try to start the DB2 database using db2start command as SIDADM user or DB2SID user, we find the below mentioned error,

"SQL6048N A communication error occurred during START or STOP DATABASE MANAGER processing"

As per the IBM mentioned procedures, we have added the entry in the .rhosts file and the db2nodes.cfg file in the /db2/db2SID path and /db2/db2SID/sqllib path. Still the issue remains the same.

Please help me in solving the error and completing the startup of the DB2 Database and the SAP Instances in Node B after the failover.

Please note - After the failover, only the IP address of the Node A is moved to Node B, as per IBM specification we have not configured the Virtual Hostname for the Nodes. On Node A the hostname is HOST-A and Node B the hostname is HOST-B. The AIX command #hostname mentions the physical hostname of corresponding hosts.

Expecting your reply at the earliest.

Thank you

Regards,

Vineeth

[Former Member · Accepted solution]

Check if following link helps you.

[http://publib.boulder.ibm.com/infocenter/db2luw/v9r5/index.jsp?topic=/com.ibm.db2.luw.messages.sql.doc/doc/msql06048n.html]

[Former Member · Accepted solution]

Dear Manoj,

Thank you for the reply.

I was able to solve the issue by following simple steps as per the DB migration methods.

The steps are mentioned below.

You need to create an alias entry in the /etc/hosts file on both systems with the following format:

ip_address short_name long_name alias

On Node A the entry looks similar to this (where Node A is your primary system):

<IP Address NODE A> <HostName Short Node A> <Fully Qualified HostName Node A> db2host (where db2host is a sample alias name)

On Node B the entry looks similar to this (where Node B is your standby system):

<IP Address NODE B> <HostName Short Node B> <Fully Qualified HostName Node B> db2host (where db2host is a sample alias name)

The ip address and domain name for both files should match those of your company network definitions for your primary and standby systems.

Now you need to ensure the hosts entry in the /etc/netsvc.conf file contains local as the first parameter. The hosts entry on both primary and standby systems would look similar to this:

hosts=local,bind

By putting local as the first parameter, this will force the system to look in the /etc/hosts file for a host name entry before going to the DNS server.

You need to change column two in the db2nodes.cfg file to a name that will be defined on both systems. Your new db2nodes.cfg file would look similar to this (Where db2host is the name to be defined on both systems):

0 db2host 0

After this

- Catalog the nodes of the cluster from Node A

- Catalog the Database from Node A

- Reset the password of SIDADM, DB2SID & SAPSID users in Node B similar to Node A

This made my cluster to fail over succesfully in to Node B.

Thank you

Regards,

Vineeth

[former_member448339 · Accepted solution]

Hi,

We are also in same situtation, we have all sap data in SAN and node1 , node2 having common service IP in AIX. We are unable to run SAP on node2 in case of failover. where as DB is running fine through node2, sap didn't started.

Please let me know what kind of cahnges we have done in profile / env to test run SAP on node2.

Thanks in advance,

Tejinder

[Former Member]

Hi Vineeth Damodar,

Please provide the below given details to analyze the problem.

Host name of Node A

Host name of Node B

After shifting the Resource to Node B what is the entry in db2nodes.cfg file.

Regards

Mohan

[Former Member]

Dear Mohan,

Please find the details mentioned below,

Host name of Node A - ARECCPD

Host name of Node B - ARECCFO

After shifting the Resource to Node B what is the entry in db2nodes.cfg file. - ARECCPD

Please help.

Thank you

Regards,

Vineeth

[Former Member]

Dear Mohan,

Please find the details mentioned below,

Host name of Node A - ARECCPD

Host name of Node B - ARECCFO

After shifting the Resource to Node B what is the entry in db2nodes.cfg file. - ARECCPD

Please help.

Thank you

Regards,

Vineeth