xiangfu0 commented on code in PR #15672:
URL: https://github.com/apache/pinot/pull/15672#discussion_r2072930849
##########
pinot-segment-local/src/main/java/org/apache/pinot/segment/local/segment/index/loader/SegmentPreProcessor.java:
##########
@@ -130,7 +130,24 @@ public void process(@Nullable SegmentOperationsThrottler
segmentOperationsThrott
if (type != StandardIndexes.forward()) {
IndexHandler handler = createHandler(type);
indexHandlers.add(handler);
- handler.updateIndices(segmentWriter);
+ try {
+ handler.updateIndices(segmentWriter);
+ } catch (IllegalStateException | UnsupportedOperationException e) {
+ // The index creation failure is expected in some cases, like when
the index is not supported for the
+ // column or segment type.
+ // E.g. non-dictionary encoded TIMESTAMP type doesn't support
range index, meanwhile dictionary encoded
Review Comment:
No, the query won't fail.
The issue is that old segments wont be able to create index and the old
segments won't be loaded.
Meanwhile the table config was updated long time back and very likely the
user who restart the cluster (SRE) are not aware this change from the real user
side(table maintainer).
So the cluster will be in the limbo state that the server won't be able to
come up unless the table is either fully disabled from the cluster, which will
impact the SLA or SRE will be able to find out the issue was due to the added
index and disable that ( it's also very hard, because the current status for
the column is updated to dictionary+ timestamp index already).
--
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]
