|
View previous topic :: View next topic
|
| Author |
Message |
abin
Active User
.jpg)
Joined: 14 Aug 2006
Posts: 198
|
|
|
|
Hi,
What is the best procedure for implementing check point re-start logic in a DL/I code.
My code updates the DL/I DB. So if the program abends in middle of a transaction I want to roll back to the last completed transaction. One transaction involves reading a DB1, Updating DB2 and inserting into DB3.
Thanks. |
|
| Back to top |
|
 |
Sandy Zimmer
Active Member
Joined: 13 Jun 2007
Posts: 826
Location: Wilmington, DE
|
|
|
|
| Is it a huge database? How long does the job run? You are running DLI or BMP? |
|
| Back to top |
|
|
abin
Active User
Joined: 14 Aug 2006
Posts: 198
|
|
|
|
| Yes it is a huge data base and it is a DLI program. The run time execution cannot be calculated now since the job is not yet in production. But on an average 2 - 3 Hrs. |
|
| Back to top |
|
|
Sandy Zimmer
Active Member
Joined: 13 Jun 2007
Posts: 826
Location: Wilmington, DE
|
|
|
|
| So it has been running for a while now and you are now just wanting to implement checkpoint-restart? Why? I also assume that you did a backup and restore procedure before considering checkpoint-restart? Have you had abends? Why? |
|
| Back to top |
|
|
abin
Active User
Joined: 14 Aug 2006
Posts: 198
|
|
|
|
actually this is a new process and is not running in production yet. So I was checking on the best procedure to follow.
The reason I am looking for re-start mechanism is because there may be abends and I do not want to re-start the entire job from start. |
|
| Back to top |
|
|
dick scherrer
Moderator Emeritus
Joined: 23 Nov 2006
Posts: 19244
Location: Inside the Matrix
|
|
|
|
Hello,
| Quote: |
| because there may be abends |
Sorry, but there should not be "regular" abends - the code needs to be abend-proof.
| Quote: |
| I do not want to re-start the entire job from start. |
Why not? If the job abends every year or 2, the complete rerun will be far more managable than some more complicated restart/recovery process.
If there is unwillingness/inability to make the code abend-proof, it is unlikely that a proper restart/recovery process would be implemented. |
|
| Back to top |
|
|
|
|
