bvaradar commented on a change in pull request #1432: [HUDI-716] Exception: Not
an Avro data file when running HoodieCleanClient.runClean
URL: https://github.com/apache/incubator-hudi/pull/1432#discussion_r396116123
##########
File path:
hudi-client/src/main/java/org/apache/hudi/client/HoodieCleanClient.java
##########
@@ -85,7 +85,11 @@ public HoodieCleanMetadata clean(String startCleanTime)
throws HoodieIOException
// If there are inflight(failed) or previously requested clean operation,
first perform them
table.getCleanTimeline().filterInflightsAndRequested().getInstants().forEach(hoodieInstant
-> {
LOG.info("There were previously unfinished cleaner operations. Finishing
Instant=" + hoodieInstant);
- runClean(table, hoodieInstant);
+ try {
+ runClean(table, hoodieInstant);
+ } catch (Exception e) {
+ LOG.warn("Failed to perform previous clean operation, instant: " +
hoodieInstant, e);
Review comment:
@lamber-ken : Can we rollback the corrupted instant seamlessly in this case ?
Can you also add a CLI (called timeline upgrade) which would read the
timeline and fix upgrade issues. For now, the CLI can just handle only known
issue - to deal with corrupted inflight clean instants. We can have users run
them before upgrade for now and also see if this can be done seamlessly.
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
