on 04-14-2009 3:17 PM
Hi,
Can someone provide me with a disaster recovery procedure for XI 3.0?
I always thought a normal online backup/restore procedure + recovery redologs + file system backup/restore would be sufficient but there's an internal discussion in our team that more is required for a consistent DR procedure.
Eg: what about XI interfaces? How preventing that a processed file is not processed twice? Etc....
We are running on Oracle 10.2.0.4, XI 3.0 SP 21
Thanks for helping me out.
BR
Bert
Hi,
The state of the messages is persisted in the database and so you should do nothing after a shutdown, the system should continue where it left if the DB can be restored to its previous state.
However, if you loose the DB or you can't restore it to the exact state it had before the crash, it is very likely you will end up either loosing messages or duplicating them.
I think that if you really need to guarantee that nothing of the sort happens, you would be better off with an XI in High Availability scenario rather than using a single instance and trying (hoping) to recover it completely.
As part of middleware product selection tests we made some rough crash scenarios and only two tools of the seven we tested came up with 0% data loss. One of them is SAP XI.
good luck!
-Sam.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi all,
Thanks for your replies.
I will soon do a test DR and hope they will run fine.
If any weird problems occur, I will share them with you on this forum.
We archive our redo logs every hour so we may loose data for max 59 minutes.
So Sam, indeed we face the risk of duplicating messages..
HA setup is not in scope at the moment.
BR
Bert
Hi,
Refer This Link.
Thanks.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
Hi,
Prepare a list of activities to be performed before and after the DR.
1.Before bringing down the actual system,you may have to stop the related jobs,stop sending Idocs to the XI system,lock the RFC users so that no more Idocs will be sent.Process all the messages and monitor them end to end before bringing down the actual system.Also,check the queues and see if they are all clear.
2.Once the DR system is up,the RFC destinations should be configured correctly to point to the target.You can then do the testing of interfaces in the DR system after activating the Communication channels.
3.Once the testing is complete and the actual system bought up,you will have to do the health check,check the RFC destinations,unlock RFC users and monitor end to end to see if the messages are flowing through successfully.
In order to avoid errors like processing the file twice,you may have to stop all the Communication channels in the Actual system before bringing them down.After the testing in the DR system is complete,you can unlock the RFC users and restart the channels so that the new messages coming in will be processed.
Hope this helps.
Thanks and Regards,
Induja
User | Count |
---|---|
90 | |
10 | |
10 | |
10 | |
7 | |
7 | |
6 | |
5 | |
4 | |
3 |
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.