divijvaidya commented on code in PR #14425:
URL: https://github.com/apache/kafka/pull/14425#discussion_r1336950457
##########
core/src/main/java/kafka/log/remote/RemoteLogManager.java:
##########
@@ -697,12 +711,22 @@ public void copyLogSegmentsToRemote(UnifiedLog log)
throws InterruptedException
this.cancel();
} catch (InterruptedException | RetriableException ex) {
throw ex;
+ } catch (CorruptIndexException ex) {
+ logger.error("Error occurred while copying log segments. Index
appeared to be corrupted for partition: {} ", topicIdPartition, ex);
+ segmentCopyFailures++;
Review Comment:
We can re-create index from the segment in case it is corrupted. This would
prevent need of any operator intervention. It will block uploading to TS for a
time duration during which index is being reconstructed but the trade-off in
favour of self-recovery is worth it in my opinion.
##########
core/src/test/java/kafka/log/remote/RemoteLogManagerTest.java:
##########
@@ -1706,6 +1758,279 @@ public RemoteLogMetadataManager
createRemoteLogMetadataManager() {
}
}
+ @Test
+ void testCorruptedTimeIndex() throws Exception {
Review Comment:
There is lot of duplicated code amongst the three tests added here. Can we
use `@ParameterizedTest` and combine them together?
##########
storage/src/main/java/org/apache/kafka/storage/internals/log/TimeIndex.java:
##########
@@ -74,13 +74,13 @@ public void sanityCheck() {
TimestampOffset entry = lastEntry();
long lastTimestamp = entry.timestamp;
long lastOffset = entry.offset;
+ if (entries() != 0 && lastOffset < baseOffset())
+ throw new CorruptIndexException("Corrupt time index found, time
index file (" + file().getAbsolutePath() + ") has "
+ + "non-zero size but the last offset is " + lastOffset + "
which is less than the first offset " + baseOffset());
Review Comment:
This is a nice change. We should fail fast by checking local state even
before calling the expensive mmap() function below.
--
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: jira-unsubscr...@kafka.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
