krummas commented on code in PR #3458:
URL: https://github.com/apache/cassandra/pull/3458#discussion_r1718056119
##########
src/java/org/apache/cassandra/tcm/RemoteProcessor.java:
##########
@@ -148,6 +148,30 @@ public ClusterMetadata fetchLogAndWait(Epoch waitFor,
Retry.Deadline retryPolicy
}
}
+ @Override
+ public LogState reconstruct(Epoch lowEpoch, Epoch highEpoch,
Retry.Deadline retryPolicy)
+ {
+ try
+ {
+ Promise<LogState> request = new AsyncPromise<>();
+ List<InetAddressAndPort> candidates = new
ArrayList<>(log.metadata().fullCMSMembers());
+ sendWithCallbackAsync(request,
+ Verb.TCM_RECONSTRUCT_EPOCH_REQ,
+ new ReconstructLogState(lowEpoch, highEpoch),
+ new CandidateIterator(candidates),
+ new
Retry.Backoff(TCMMetrics.instance.fetchLogRetries));
+ return request.get(retryPolicy.remainingNanos(),
TimeUnit.NANOSECONDS);
Review Comment:
I don't know how accord will actually use this, but I imagine this to be
quite "bursty" - like right after we join, there would be many transactions
using older epochs that need reconstruction, so a debounce + store the result
locally would reduce the load on the CMS
But maybe thats not how it will be used?
What does "single-shot feature" mean here?
--
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]
