Best practice recmmendations for backups

Question

Hi,
I wanted advise from senior basis consultants on backup strategy.
Here is a summary of our environment:
We have 6 SAP servers, with more than 1 SAP system installed per server.
The SAP systems include CRM, SRM, XI, Portal, Solution Mgr, ERP , R/3 , BW. Many of these have java stack enabled with integrated ITS etc.
The systems are internal and used for sales demos, not productive
On an average 1 to 2 business days of functional work is done on most of these systems per week.
Admin changes such as support packs, kernel patches, minor config changes to config files on disk ( e.g. for IPC, TREX etc. ) are done on a routine basis.
It took us weeks to get some of these systems installed and all the  issues ironed out. So re-installing all applications from scratch and restoring the database is not the best option for us.
We are willing to live with losing a week's worth of data in case of a failure.
We currently have OS licenses for Veritas, but no MSSQL server agent licenses.
Right now we do 1 offline OS backup every weekend. Database online backups are done once a week to disk and txn logs daily to disk. I am planning to drop the db backups though. I know SAP's recommendations for productive systems, but that is overkill for us. Alternatively any SAP  documents that describe this?
Thanks in advance,
Sanket

Answer 1

Hi, I came across this message while searching for something else.  I also was asked to find the "Best Practices" for backup strategy in the past.  I found nothing from SAP.
I get the impression that SAP doesn't produce such a thing.  When I created the backup strategy for my client, I think I understand why this might be the case.
I think you can find documentation on what is technically required to make sure that you can restore/recover a database.  I don't even think I have seen documentation from SAP on what you have to preserve at the OS level to restore a system.
But, technical requirements for recovery are not the same as a backup strategy.  When you start talking about this subject, you have to know some things about the organization, about how much it costs to be down, and how much you are willing to spend to prevent downtime and data loss. 
It depends on the size of the organization.  You have to know whether the Basis person also do the DB admin work, the OS admin work, the "tape hanging" work, etc.  In the SMB world, it sometimes is the case that one person does all the work, so there is no coordination to be done.  But you still have to coordinate "backup coverage" for the activity, in the event of absence/separation of the primary resource.
You also need to test the backup strategy, too.
Based on these inputs, and others, you synthesize a strategy that guarantees that your organization is covered by the backups.