Stefan Egli created SLING-3613:
----------------------------------
Summary: [discovery] verify heartbeat config at connect time,
reject if diffs
Key: SLING-3613
URL: https://issues.apache.org/jira/browse/SLING-3613
Project: Sling
Issue Type: Bug
Components: Extensions
Affects Versions: Discovery Impl 1.0.8
Reporter: Stefan Egli
Discovery.impl's heartbeat mechanism relies on one crucial detail: both ends of
a topology connector must have the same configuration wrt interval and timeout.
With 'intra-cluster' heartbeats, going to the shared repository, the same is
true, but this is guaranteed by the fact that the corresponding OSGi config is
also shared by all the instances. With 'inter-cluster' heartbeats though - so
called 'connectors' - you could have different instances configured
differently. If that's the case, you could end up having an instance ping too
slowly and the peer timing out that instance constantly (if its heartbeat
timeout is smaller than the incoming heartbeat interval).
The safest way to handle this is to double-check the config parameters at
connector-connect time, and reject connectors if they have different config
parameters. That way this will get noticed by administrators way faster than by
simply logging an error.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
