Setup of EP 6.0 SP10 cluster with 2 nodes

[Former Member]

Hi,

We are in the process of installing EP 6.0 SP10 (from the Netweaver '04 SR1 CDs) on a linux RHEL 3.2/ORACLE 9.2.0.5 (all 32-bit) platorm.

We have two nodes - node A and node B. (Note- FWIW- the oracle database will run on a separate linux server)

Node A will run:

1. SCS

2. Dispatcher

3. server processes

Node B will run:

2. Dispatcher

3. server processes

Load balancing between the two nodes will be handled externally via a 3rd party hardware solution.

The $1,000,000 question- What can we do about redundancy of the SCS services?

We would like something similar to the "backup state controller" install option that was available for the older EP 6.0 SP2 architecture.

Any constructive advise or pointers to other documentation, OSS notes, etc. will be much appreciated!

Sincerely,

Steven McElwee, Duke University

[jpolus · Accepted solution]

Hi Steven,

This falls into to NetWeaver High Avaibility. Check note 803018.

John

[Former Member · Accepted solution]

This note implies the possibility that consultants and/or customers might have already reached a solution, or, perhaps, might be able to shed additional information about this.

So, to anyone reading this posting thread, any ideas?

Sincerely,

Steven McElwee, Duke University

[Former Member · Accepted solution]

Note 803018 = £/$.

I am doing the same setup on MS SQL Server on a distributed scenario.

Try the scenario on your own, obviously on a (sandpit) dual server.

James

[Former Member · Accepted solution]

We are in the process of building the entire landscape. I am just in the process of building the development system. Unlike q&a and production, the development server will consist of 1 linux server running the database in addition to SCS, dispatcher and server services.

Q&A and Production will each consist of Node 1 and Node 2 as described above along with another linux server that runs the oracle database.

I've already got a sandbox system, so i'm confident of building all this. I just don't know what to do about the SCS replication redundancy issue for the Q&A and Production nodes.

In fact, i've done some research at services.sap.com/ha and, aside from finding a massive 254 page document entitled "Sap High Availability" which does not directly address the portal product, haven't found a good starting point for this. Right now, I would even settle for a good technical document that explains more clearly the SCS architecture. IMHO, "Technical Infrastructure of SAP Enterprise Portal 6.0 on SAP Web AS 6.40" guide, while providing a fairly good overview of the portal architecture, does not provide enough lower-level detail - most especially about the SCS services.

So how about it, are there any folks out there willing to share their hard-earned experiences with the rest of us "would-be" pioneers?

Here's hoping ...

Steven

[Former Member · Accepted solution]

Steven,

Drop me your email and i can send you some docs. What I would do is setup a separate SCS environment (planned for NW04/SR1) most probably within a switch-over environment.

James

jlorenzana@virgin.net

[Former Member · Accepted solution]

Sorry,

it is j.lorenzana@virgin.net

Regards,

James

[Former Member · Accepted solution]

James,

thanks for your response. My email address (apologies for not including it previously!) is <steven.mcelwee@duke.edu>.

Sincerely,

Steven

[Former Member · Accepted solution]

James,

I just rewarded you with a green star!

John,

I just rewarded you with a yellow (or is it gold?) star?

Steven

[Benny · Accepted solution]

May I state here, that Note 803018 does NOT mention any service that inherits any additional cost!

Thisd service is free of charge. Not using it means that SAP does not guaranty that your software will run as expected!

Regards,

Benny

[Former Member · Accepted solution]

Benny,

The note (803018), does that explicitly that it is charged consulting.

Steven, send me email, I have node A+ node B working. I just haven't had time to document.

Regards,

James

j.lorenzana@virgin.net

[Former Member · Accepted solution]

Hello James,

We too are in similar situation with multi node setup. Can you send me the docs you sent to Steven? It would be greatly appreciated.

Thanks,

Dan

dschultz@reliant.com

[Former Member · Accepted solution]

Dan,

The docs which James sent me are a pdf slide presentation from a class at a conference. They give good detail of the portal architecture and are available here at the sdn.sap.com web site by:

on SDN (Home) search box, enter;

PRTL203

It is the second result (PDF)

Good luck,

Steven

[Former Member · Accepted solution]

James,

My email address is <steven.mcelwee@duke.edu>. I am looking forward to seeing how you got node A and node B working!

Thanks for being willing to share this information.

Sincerely,

Steven

steven.mcelwee@duke.edu

[Benny · Accepted solution]

James,

the registration and evaluation process is NOT. I know and guarantee for it as I'm working for this project.

Of course getting consulting to make up an architecture is not free.

Regards,

Benny

[Former Member · Accepted solution]

Benny,

Thanks. Perhaps, there should be a product offering *.ppt or whitepaper of what is free and what is not besides the note.

Regards,

James

[Benny · Accepted solution]

That is in the works and will be released about next week.

I'll take care to communicate this then.

Regards,

Benny

[Former Member · Accepted solution]

James hopes to have the documentation for how he accomplished this sometime at the end of March.

Sincerely,

Steven Mcelwee, Duke University

[Benny · Accepted solution]

Took a little bit longer, but now we have released the new WebAs 6.40 switchover guide that has the basics in a pdf document. Please got to service marketplace /ha to get the doc.

Regards,

Benny