github-actions[bot] commented on code in PR #65553:
URL: https://github.com/apache/doris/pull/65553#discussion_r3575825421
##########
fe/fe-core/src/main/java/org/apache/doris/load/routineload/RoutineLoadJob.java:
##########
@@ -1712,6 +1724,7 @@ public List<String> getShowInfo() {
row.add(userIdentity.getQualifiedUser());
row.add(comment);
row.add(getClusterInfo());
+ row.add(getFirstErrorMsg());
return row;
} finally {
readUnlock();
Review Comment:
This timestamps the diagnostic when FE processes the attachment, but the URL
lifetime is controlled by BE error-log file age. A routine-load job can legally
set `max_batch_interval` above the default 48-hour BE
`load_error_log_reserve_hours`; if the first bad row appears early and the
transaction callback arrives after BE cleanup, FE records a fresh timestamp
here and keeps showing a dead URL/message for another retention window. Replay
has the same problem because old attachments are repopulated with the replay
time. Please timestamp this from a value tied to the BE log lifetime, such as
task/error-log creation time or transaction finish time, and skip or expire
entries whose backing log may already be gone.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
