Addressing truncated change log or rollback detected errors
Problem
Baselogindexing occurs when a TRS is added to Lifecycle Query Engine.Baselogindexing fetches the current set of resources from the application.TRS Changelogprocessing occurs afterBaselogindexing when Lifecycle Query Engine frequently checks with the TRS for any changes, and updates them while indexing.
During changelog processing, Lifecycle Query Engine validates the TRS by
looking at the last processed ChangeLogEvent. If the last processed
ChangeLogEvent from the application matches the last processed
ChangeLogEvent
Lifecycle Query Engine indexed, then the change logs are processed. If the last processed
ChangeLogEvent does not match, then the TRS validation fails with a
TruncatedChangeLogException.
These failures typically happen because an application has reset its
ChangeLogEvents and they no longer match the information that is indexed through
Lifecycle Query Engine. The failures can happen if the application TRS is rebased or reset.
The only way to recover from a RollbackDetectedException or a
TruncatedChangeLogException error is to re-index the application TRS.
Solution
When a data source fails with TruncatedChangeLogException, it's possible that
there's no actual truncation instead Lifecycle Query Engine failed to get the next page due
to some transient condition. Therefore, on the data source page in Lifecycle Query Engine,
click Force update. The Lifecycle Query Engine checks and processes
the change log immediately in the TRS feed. If the transient condition is cleared, the data source
resumes the normal operation.
RollbackDetectedException or a TruncatedChangeLogException error.
It is expected that after an Engineering Requirements Management DOORS® Next rebase, you must reindex the TRS in
Lifecycle Query Engine. - If during indexing, you encounter the
RollbackDetectedExceptionorTruncatedChangeLogExceptionerror, and it has many skipped resources, do not try to reindex the data provider. Investigate and address the skipped resource issue. When the resource issues are resolved, reindex the data provider. - If you encounter that the TRS event is present in the base, and hence it is present in TRS, then
reindexing fixes it. If the TRS event disappears from the
Changelog, then it is not present in the base. It means that the TRS event is not present in the TRS feed and hence reindexing does not help. - If DOORS Next or Engineering Test Management resources has a data gap and the Data providers are displayed up to date in Lifecycle Query Engine, then the TRS validation must be done. TRS validation can be first done on the Data provider side (DNG or Engineering Test Management application) and then on the Data consumer side (Lifecycle Query Engine application). Validation is faster than re indexing. If validation does not solve the problem, only then opt for re indexing.