MaxDB databases in HP_UX Service Guard Cluster packages - Anyone?

[Former Member]

Hi MaxDB People

Here is my second post/question. Hopefully somebody can help me out here.

I've installed a MaxDB database, and performed a Database migration from Oracle to create a SAP Java ADS instance called AQ3.

We use HP Service guard clusters, to allow us to move instances around between server nodes. Because if this, I have configured the file systems for the database and software as follows:

Node 1 (where the original install took place)

<Local to the server node>

/sapdb

/sapdb/data

/sapdb/programs

/sapdb/clients

<Installed as a movable package called 'sapaq3' and mounted in bold>

/sapdb/AQ3/  <and all subsequent directories>

/sapdb/client/AQ3 <and all subsequent directories>

/usr/sap/AQ3

/sapmnt/AQ3

On the original server node this instance runs just fine.

Node 2

I installed the MaxDB software again locally

/sapdb

/sapdb/data

/sapdb/programs

/sapdb/clients

I then moved the package 'sapaq3' to run on node 2

/sapdb/AQ3/  <and all subsequent directories>

/sapdb/client/AQ3 <and all subsequent directories>

/usr/sap/AQ3

/sapmnt/AQ3

I started the x_server from /sapdb/AQ3/db/bin

When I try to start the database however, I am getting the following error message:

dbmcli -d AQ3 -u control,xxxxxxxx db_online

Error! Connection failed to node (local) for database AQ3:

database instance not found

I'm thinking there must be some need for a database registration on node 2, so that the central software knows about the database that has been previously installed elsewhere. I tried the sdbreg command but I am still getting the error.

Does anyone have any idea what the issue could be?

Thanks everyone.

Bill Dedman

[Former Member · Accepted solution]

Hi Bill,

Did you mounted the saplog & sapdata as well on the 2nd node as part of the cluster config as this will be required.

and do verify your config file is updated on the secondary node and the hostname within the config is your virtual name ( in this case i assue sapaq3).

Also you may need to check the volume configurations on the 2nd node for the correct mapping of /dev to your file systems. you may need to attach the sap log & data volumes  again on 2ndary node again.

Regards,

ram

[Former Member · Accepted solution]

Hi Ram

Thanks for the response. I should have added, this is part of a database migration test, from an oracle instance, so I am reusing the hostname, SID, and Instance numbers from the old Oracle install of AQ3.

So, yes, I have got the sapdata and saplog directories mounted on  further lvols beneath /sapdb/AQ3 as part of the package. The package is not the problem I think, as it happily starts on all four nodes of my cluster.

I'm still looking for that resolution that will enable me to start the database on a second node, sadly.

Thanks anyway.

Bill

[Former Member · Accepted solution]

Also do refer to below Document and the oss note

http://h20566.www2.hpe.com/hpsc/doc/public/display?sp4ts.oid=4162088&docId=emr_na-c03253067&docLocal...

1765651 - Support details for HP Serviceguard

Regards,

Ram

[Former Member · Accepted solution]

Hi Bill,

Can you provide below details for the Maxdb instance on the second node.

for dbms details:-

sdbregview -l

For instance details:-

dbmcli -s inst_enum

dbmcli -s db_enum

And run ./sdbverify on your second node.

This will provide us on more data points on the second node instance.

And can you provide your inputs for the setup based on oss note 803452 - MaxDB directories in UNIX clusters for failover.

whether all required configuration is done.

Regards,

Ram

[thorsten_zielke · Accepted solution]

Hi Bill,

if this is about a SAP software installation, I would suggest that you create a SAP support ticket.

In general, for Unix systems we rely on our hardware vendor partners to offer support and assistance for the different cluster solutions.

Kind regards,
Thorsten

[Former Member]

Hi All

Thanks for your help in this issue. I just wanted to let you know that I've fixed the problem in the following way:

After installing the AS Java instance on the first node, and the SAPDB software on the other three nodes of the cluster, I was able to restart the Instance after copying the files Databases.ini and Installations.ini in /sapdb/data/config from the original cluster node, out to the other nodes. It was just a question of synching the config info around the cluster, as I suspected.

Thanks again for your help. Much appreciated.

Bill Dedman

[Former Member]

Hi Bill,

Good to hear that it was resolved. Request you to close the thread by making it answered.

Regards,

Ram