aweisberg commented on code in PR #4111:
URL: https://github.com/apache/cassandra/pull/4111#discussion_r2122077048
##########
src/java/org/apache/cassandra/exceptions/RequestTimeoutException.java:
##########
@@ -40,4 +40,13 @@ protected RequestTimeoutException(ExceptionCode code,
ConsistencyLevel consisten
this.received = received;
this.blockFor = blockFor;
}
+
+ // TODO (review): Was there an intentional choice here that
RequestTimeoutException doesn't wrap its causes for brevity?
Review Comment:
Reviewing this again I think it's not an issue for a bootstrapped range and
it's not an issue for a range that Accord is aware of but doesn't actually run
transactions for.
Where I am not clear is is the case where it's a range that is already owned
and thus won't be bootstrapped?
```
Map<Boolean, Ranges> partitioned =
addRanges.partitioningBy(range -> shouldBootstrap(node, prev.global,
newLocalTopology, range));
if (partitioned.containsKey(false))
bootstrapUpdates.add(shard.store.initialise(epoch,
partitioned.get(false)));
```
`CommandStore.initialize` uses `bootstrapBeganAt` and does
`unsafeSetBootstrapBeganAt` but doesn't update max conflicts.
`loadBootstrapBeganAt` does update `maxConflicts`. The other issue is it
doesn't seem like `bootstrapBeganAt` would have be anything other than
`emptyBootstrapBeganAt()`.
--
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]
