the question about the HA installation on ECC6.0

[Former Member]

Hi Experts,

We are about to implement a project with HA environment on the ECC6.0 in the near future, which is just about the ABAP stack. After reading the Installating Guide, I stil have several questions related to the procedures of HA installation.

In the guide document, I got the following steps to process for realizing the HA of ECC6.0:

1. Run SAPinst to install the central services instance (ASCS) using the virtual host name on the primary cluster node, host A.

2. Prepare the standby node, host B, making sure that it meets the hardware and software requirements and it has all the necessary file systems, mount points, and (if required) Network File System (NFS), as described in Preparing for Switchover .

3. Set up the user environment on the standby node, host B. For more information, see Creating Operating System Users and Groups Manually. Make sure thatyou use the same user and group IDs as on the primary node. Create the home directories of users and copy all files from the home directory of the primary node.

4. Configure the switchover software and test that switchover functions correctly.

5. Install the database instance on the primary node, host A.

6. Install the central instance with SAPinst on the primary node, host A.

7.If required, install additional dialog instances with SAPinst to replicate the SAP system services that are not a SPOF. These nodes do not need to be part of the cluster.

My Question is that does standby node(host B in above context) need to install the ASCS, database instance and Central Instance?

If host B does not need to install the database instance, how about the whole system would be when the primary cluster node (Host A in above context) totally crash, such as power failure.

[Former Member · Accepted solution]

Hi Rajesh,

The last question.

Should I install the Java Runtime Environment on the local disk in the primary node and standby node separately? I mean both in the ECC 6.0 and PI 7.0 using only CI instance way to realize the HA solution.

Thank you very much, Rajesh.

Best Regards,

Rongjie

[Former Member · Accepted solution]

Hi Rong,

JAVA (JDK) installation would be local...

You need to install same JDK Version on all nodes, as the location of JDK installation is usually "/opt/jdk1.4" directory which is not a part of shared directories..

All the best !...

Regards.

Rajesh Narkhede

[Former Member]

Hi Rajesh,

Ok,I got it. Can I use the same way to install the PI 7.0 with the HA solution? I am not very sure about it becaust it is the java stack.

Thanks and Best Regards,

Rongjie

[Former Member]

Hi Rong,

Can I use the same way to install the PI 7.0 with the HA solution?

Sure, you can install it like that way...

(Java instance will be installed in /usr/sap/<JAVA_Instance> and /sapmnt/<JAVA_Instance>, so there is no problem in using this technique with JAVA instance also...)

Main logic for this is, you should keep all those directories on SAN disk which is being created by SAP installation and of course Oracle directories too..

So when you do switchover, all those Application + Database data will be available on second node.

Regards.

Rajesh Narkhede

[Former Member]

Hi Rajesh,

Ok,I got it. Can I use the same way to install the PI 7.0 with the HA solution?

Thanks and Best Regards,

Rongjie

[Former Member]

Hi,

Is my installation plan right in such way?

Thanks and Best Regard,

Rongjie

[Former Member]

Hi Rajesh,

I really appreciate your advice.

But due to some reasons, the customer ask to use the only CI instance way which also called Central System installing method in SAP installation document to establish the HA system. Let me short also.

Using this way to make this kind of HA working, here is the next steps I should follow:

1. using Central System way to install the primary node as normal way of installation.(make sure that using virtual name to recognize the host, and /sap/usr,/sapmnt,/oracle directory on the SAN storage)

2. configurating the standby node, including create user, copying related home files from primary node to standby node as well as mount /sap/usr, /sapmnt and /oracle directories on SAN storage to standby node.etc.

3. test the switchover functions.

Is that right if using Central System way to make the HA running?

[Former Member]

Hi Rong,

If your customer is asking for CI instance only then your approach is correct.

1. You can install First node as normal installation.

2. Switchover the cluster.

3. /usr/sap, /sapmnt, /oracle must be mounted on Second Node.

4. Make the user environment same on both node (username, profiles)

5. Test the by starting up the SAP.

Regards.

Rajesh Narkhede

[Former Member]

Hi Rajesh,

Thanks for your further explanation.

I have already gone throuth the SAP help relating to the Standalone Enqueue Server because you mentioned it on preivious messages. But in our case, we did not plan on installing a Standalone Enqueue Server individually as there is no requirement for replica of the lock table. Is it necessary to must use the ASCS instance to install the HA? I mean, theoretically, whether we can use Central System installing method to implement the HA solution or not just in our case.

[Former Member]

Hi,

--- Short Answer.

If you want to get the benefits of HA setup, you should go with ASCS instance.

Practically, It may be possible to use only CI instance, but then there is no meaning of implementing HA setup.

I would suggest, follow the standard procedure for HA setup as recommended by SAP. Then you will realise, how actually HA works. and of course about its benefits also...

Regards.

Rajesh Narkhede

[Former Member]

Great Information Rajesh.

I also benefit lot from it .

Thanks and Regards,

Arun Kumar

[Former Member]

Hi Rajesh,

Thanks for your further explanation.

Using the this method to realize the HA architecture, I think we may do not have to use the sequence of the first ASCS and then Database instance just as installation guide describe for the HA. Just using the Central System installing method on the primary node first as we usually do in the normal system, it will also can do the everyting that you mentioned on last several messages on the standby node. I considered that the two installing methods which is ASCS or Central System is no difference if using the method we talk about above.

[Former Member]

Hi Rong,

No !...

ASCS instance is required in HA setup, as it inclueds enque replication server.

I think for such doubts you should go through the SAP help..

Try this below link:

[http://help.sap.com/erp2005_ehp_04/helpdata/EN/6d/e3a03cc24efd1de10000000a114084/frameset.htm|http://help.sap.com/erp2005_ehp_04/helpdata/EN/6d/e3a03cc24efd1de10000000a114084/frameset.htm]

Regards.

Rajesh Narkhede

[Former Member]

Hi Mark,

Since standby node install nothing(neither SAP nor Oracle), I should copy the environment files of Oracle and sap<side> users from the primary node to the standby node in order to bring up the Oracle and SAP system on standby node. In the installation guide, I can obtain the relative information of SAP but the information about Oracle is not mentioned in that guide. How can get the oracle working on the standby node when switchover happened.

Thanks

[Former Member]

Hi Rong,

Since standby node install nothing(neither SAP nor Oracle)

Keep in mind, even if you are installing it from Node A, you won't need to install it on Node B. Because, installation data will be available on Node B as it is installed on SHARED STORAGE (SAN DISK). The same will be mounted on another Node after switchover. So the installed data will not lost. Those all mount point will be available on only ACTIVE Node, no matter whether it is Node A or Node B.

How can we get the oracle working on the standby node when switchover happened.

When Switchover happened your /oracle mount points will be mounted on Node B.

To start the oracle you will need following:

1. ora<sid> operating system user.

2. Oracle related environment settings (ora<sid> users profiles)

3. ORACLE_HOME directory (ie. /oracle/<SID>/102_64)

4. Oracle Database files (/oracle/<SID>/sapdata<n>, /oracle<SID</mirrlog/<n>, /oracle/<SID>origlog<n> etc.)

5. Root oracle volume (i.e. /oracle)

Above things are sufficient to bring up the oracle database.

Same thing is for SAP instance. (<sid>adm user and SAP mount points. For example: /usr/sap, /sapmnt/)

Regards.

Rajesh Narkhede

[Former Member]

Hi Rajesh,

Thanks for your kindly explanation.

As you said on the last reply, when primary node failed or crashed, only users and thier enviornment setting will be lost. My question is that how about the consistency of oracle at this circumstance. Due to the mechanism of SGA of Oracle, the memory area for Oracle like database buffer cache or Redo log buffer cache area can not synchronize with its own database in time because of the sudden failover of primary nodes. Is there any problem when second node bring up both the Oracle and SAP system?

[Former Member]

Hi Rong,

During a system crash of the primary node, indeed some Oracle data (manly buffered data) can be loss d. The most important data in order to be able to recover your database is located in the online redo logs. OriglogA/B.

I assume that these filesystems are part of your cluster and will be mounted on the secondary node. So after a crash when you startup Oracle on the secondary node, you probably have to recover the last part of your database. Oracle will notice this and give you a warning regarding this. Since Oracle is solid as a rock, you will be fine....

What you also could do, is instead of failing over with your filesystems, is using an Oracle standby database. However, in case of a crash, I think the first solution would be better, since you can go back to the last point in time.

Kind regards,

Mark

[Former Member]

Hi Rong,

I would try to explain it in simple words...

My Question is that does standby node(host B in above context) need to install the ASCS, 
database instance and Central Instance?
If host B does not need to install the database instance, how about the whole system would be when 
the primary cluster node (Host A in above context) totally crash, such as power failure.

1. You don't need to install ASCS on Node B. You are installing it using a VIRTUAL HOSTNAME which represent cluster not individual node. VIRTUAL HOSTNAME is assigned to cluster package, so whichever Node is the owner of the package, will have the VIRTUAL HOSTNAME. (it will switch with cluster switchover)

2. It is actually a cluster package configuration magic. When Node A is active, cluster package owner is Node A. So all mount points (which is on SAN disk) is mounted on Node A. When you switchover the cluster, those packages will be mounted on Node B.

Some time single cluster package is used (which includes mount points for SAP instance + Database directories). You can also use 2 cluster package seperating SAP and Database directory structure.

Only OS related directories should be on servers local disk. All other application related mount points should be on SAN disk which is configured in "Cluster Package". (For example /sapmnt, /usr/sap, /oracle etc.)

You only need identical users and their enviornment settings on both Nodes.

In simple words, When primary node fails or crashed only users and thier enviornment setting will be lost. On second node, because of identical users and their profiles, same settings will be available to bring up the SAP system. Your all SAP and Database data is intact as it is on SAN Disk.

I hope, your confusion is cleared now...

Regards.

Rajesh Narkhede

[Former Member]

Hi, Mark:

If nothing is installed on the failover node, then how the failover node can take over the database instance and central instance.

I mean if the failover node just simply use the oracle binary where is in the oracle home directory located on the SAN to restart the database instance and central instance as primary node crashed.

Thanks

[Former Member]

Hi Rong,

Of course you need to execute some activities on the failover node, in order to get it working. But as far as I know, these are manual steps, not executed by sapinst. (however, I have never executed these steps with sapinst...)

I would expect that the Oracle software is part of the clusterd storage, so this will be mounted on the failover node automatically. If not, you indeed have to install the Oracle software manually.

Also the environment files of the Oracle and sap<sid> users need to be copied from the primary node to the secondary node.

Kind regards,

Mark

[Former Member]

Hi, Mark:

Thanks for your reply.

I don't think this scenario that provided in the installation guide is for the usuage of enqueue replication server on Node B. As you saw above steps, there are nothing installed in Node B, which is also my question. Just Step 2 and 3 mentioned relative NFS and user environment setting etc on the Node B.

Doesn's Node B should install any instance at all.

[Former Member]

By default nothing is installed on the failover node. You can choose to install an application server on this node, end- or an replication server. Check the following link for ERS, in case you are interested: http://help.sap.com/saphelp_nw70ehp1/helpdata/en/47/e023f3bf423c83e10000000a42189c/frameset.htm

Kind regards,

Mark

[Former Member]

Hello Rong,

Node B will be used in your case for the enqueue replication server and after a failover (due to whatever reason), the central instance and database instance will be switched to node B.

You could install an application server on node B, to make use of this additional capacity. In case of a failover, you have to run with less capacity.

Hope this helps..

Kind regards,

Mark Dijsselbloem