Issue
The capture job fails with the below error:
F_JZ0A42: The online transaction log file has been reused already, and no matching transaction log backup file is found
Why is the capture job trying to go back till the start of the last open transaction detected, and re-read the changes?
Environment
Source: SAP HANA
Answer
- Capture time reset
- Full refresh on all the tables impacted
- Extension of the log retention period on the source side, for allowing Local Data Processing recovery in case of any ad-hoc events like this.
Cause
Local Data Processing has a feature to create checkpoints, preventing the capture job from reading the already processed logfiles, but unfortunately, it is not available for HANA sources.
After a small disconnection or an event when the HANA source isn't available, Local Data Processing capture job needs to go back till the start of the last open transaction detected, and re-read the changes which can cause long-running transactions and growing latency easily.
Many times those transaction logs have already been purged/archived/moved from the location, so the job ends up in error. Having the log retention period strict is increasing the chance for the capture job to end up in an error like the above.