Stefan Egli created SLING-5256: ---------------------------------- Summary: change in localClusterSyncTokenId should always trigger a TOPOLOGY_CHANGED Key: SLING-5256 URL: https://issues.apache.org/jira/browse/SLING-5256 Project: Sling Issue Type: Bug Components: Extensions Affects Versions: Discovery Oak 1.0.2, Discovery Base 1.0.2, Discovery Commons 1.0.2 Reporter: Stefan Egli Assignee: Stefan Egli Priority: Minor
Normally when a topology changes this is either because an instance joined or left (or because properties change, but that's another topic). And whenever an instance joins or leaves this is triggering a TOPOLOGY_CHANGED as expected. However there could be the unlikely scenario that from a state (a) an instance joins creating state (b), then leaves relatively quickly again resulting in state (c) - and one other instance in the cluster happens to be so busy that it 'misses' the intermediate state (b) - which means it will compare state (a) with state (c) which it could see as being equal since, well, that it almost is. But to be correct, the above comparison between (a) and (c) should still trigger a TOPOLOGY_CHANGED event. And that is already possible since (a) and (c) have a differing {{localClusterSyncTokenId}} (by definition). Long story short: the code should treat two TopologyViews with differing {{localClusterSyncTokenId}} as different (thus automatically causing a TOPOLOGY_CHANGED as a result) -- This message was sent by Atlassian JIRA (v6.3.4#6332)