dcapwell commented on code in PR #3416:
URL: https://github.com/apache/cassandra/pull/3416#discussion_r1683343368
##########
src/java/org/apache/cassandra/service/accord/AccordService.java:
##########
@@ -358,6 +464,79 @@ public synchronized void startup()
state = State.STARTED;
}
+ private static boolean isSyncComplete(Ranges ranges)
+ {
+ for (Range range : ranges)
+ {
+ TokenRange tr = (TokenRange) range;
+ if (!tr.isFullRange())
+ return false;
+ }
+ return true;
+ }
+
+ private List<ClusterMetadata> discoverHistoric(Node node,
ClusterMetadataService cms, OptionalLong optMaxEpoch)
+ {
+ ClusterMetadata current = cms.metadata();
+ Topology topology = AccordTopology.createAccordTopology(current);
+ Ranges localRanges = topology.rangesForNode(node.id());
+ if (!localRanges.isEmpty()) // already joined, nothing to see here
+ return Collections.emptyList();
+
+ Map<InetAddressAndPort, Set<TokenRange>> peers = new HashMap<>();
+ for (KeyspaceMetadata keyspace : current.schema.getKeyspaces())
+ {
+ List<TableMetadata> tables =
keyspace.tables.stream().filter(TableMetadata::requiresAccordSupport).collect(Collectors.toList());
+ if (tables.isEmpty())
+ continue;
+ DataPlacement placement =
current.placements.get(keyspace.params.replication);
Review Comment:
thought I replied to this... when the node starts up we need to know how
placement changes, this currently does show the nodes leaving. I think the
only way to work around `whenSettled` is to diff write from read, but that
feels a bit brittle for me as it makes assumptions on *why* we are in that
state.
--
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]
