Merge branch 'cassandra-3.11' into trunk
Project: http://git-wip-us.apache.org/repos/asf/cassandra/repo Commit: http://git-wip-us.apache.org/repos/asf/cassandra/commit/50ba850b Tree: http://git-wip-us.apache.org/repos/asf/cassandra/tree/50ba850b Diff: http://git-wip-us.apache.org/repos/asf/cassandra/diff/50ba850b Branch: refs/heads/trunk Commit: 50ba850beb573ae4aa62ac46e645ce48413815ce Parents: b687641 4d24958 Author: Paulo Motta <pa...@apache.org> Authored: Thu Apr 20 10:40:22 2017 -0300 Committer: Paulo Motta <pa...@apache.org> Committed: Thu Apr 20 10:48:37 2017 -0300 ---------------------------------------------------------------------- CHANGES.txt | 1 + .../cassandra/service/ActiveRepairService.java | 21 ++++++++++++-------- 2 files changed, 14 insertions(+), 8 deletions(-) ---------------------------------------------------------------------- http://git-wip-us.apache.org/repos/asf/cassandra/blob/50ba850b/CHANGES.txt ---------------------------------------------------------------------- diff --cc CHANGES.txt index db863e1,08df2dd..2710060 --- a/CHANGES.txt +++ b/CHANGES.txt @@@ -80,7 -24,9 +80,8 @@@ * NoReplicationTokenAllocator should work with zero replication factor (CASSANDRA-12983) * Address message coalescing regression (CASSANDRA-12676) * Delete illegal character from StandardTokenizerImpl.jflex (CASSANDRA-13417) - * Fix cqlsh automatic protocol downgrade regression (CASSANDRA-13307) Merged from 3.0: + * Fail repair if insufficient responses received (CASSANDRA-13397) * Fix SSTableLoader fail when the loaded table contains dropped columns (CASSANDRA-13276) * Avoid name clashes in CassandraIndexTest (CASSANDRA-13427) * Handling partially written hint files (CASSANDRA-12728) http://git-wip-us.apache.org/repos/asf/cassandra/blob/50ba850b/src/java/org/apache/cassandra/service/ActiveRepairService.java ---------------------------------------------------------------------- diff --cc src/java/org/apache/cassandra/service/ActiveRepairService.java index 381af48,fea690e..c03c470 --- a/src/java/org/apache/cassandra/service/ActiveRepairService.java +++ b/src/java/org/apache/cassandra/service/ActiveRepairService.java @@@ -358,14 -324,17 +358,16 @@@ public class ActiveRepairService implem } else { - status.set(false); - failedNodes.add(neighbour.getHostAddress()); - prepareLatch.countDown(); + // bailout early to avoid potentially waiting for a long time. + failRepair(parentRepairSession, "Endpoint not alive: " + neighbour); } } - try { - prepareLatch.await(DatabaseDescriptor.getRpcTimeout(), TimeUnit.MILLISECONDS); + // Failed repair is expensive so we wait for longer time. + if (!prepareLatch.await(1, TimeUnit.HOURS)) { + failRepair(parentRepairSession, "Did not get replies from all endpoints."); + } } catch (InterruptedException e) { @@@ -382,9 -349,13 +382,14 @@@ return parentRepairSession; } + private void failRepair(UUID parentRepairSession, String errorMsg) { + removeParentRepairSession(parentRepairSession); + throw new RuntimeException(errorMsg); + } + - public void registerParentRepairSession(UUID parentRepairSession, InetAddress coordinator, List<ColumnFamilyStore> columnFamilyStores, Collection<Range<Token>> ranges, boolean isIncremental, long timestamp, boolean isGlobal) + public void registerParentRepairSession(UUID parentRepairSession, InetAddress coordinator, List<ColumnFamilyStore> columnFamilyStores, Collection<Range<Token>> ranges, boolean isIncremental, long repairedAt, boolean isGlobal) { + assert isIncremental || repairedAt == ActiveRepairService.UNREPAIRED_SSTABLE; if (!registeredForEndpointChanges) { Gossiper.instance.register(this);
