[jboss-jira] [JBoss JIRA] (JGRP-1523) FD_ALL does not unsuspect on heartbeat

[ https://issues.jboss.org/browse/JGRP-1523?page=com.atlassian.jira.plugin.... ] Jan Boehm commented on JGRP-1523: --------------------------------- As I understand FD_ALL.suspect() it only propagates the suspects if it believes to be the (new?) coordinator (if local_add.equals(first)). Supposed you have a member list with the following suspect states [old_suspect, new_suspect, me, ...] while suspect() is called: - Old_suspect indicates a wrongly suspected node that was retained in the suspect list (the suspicion was not propagated, because new_suspect was not suspect at the time and thus suspect() did not propagate the suspicion). - New_suspect is a new suspect, passed to the current call of suspect(). - Me is the local node. In this case the local node would start verifying its suspicion despite the fact, that old_suspect may be perfectly reachable now and according to the "first" check should be the new coordinator and thus responsible for checking new_suspect. Contrast this with the case in which old_suspect was never suspected: Now the local node would not propagate the suspicion about new_suspect. I am not sure if this inconsistency causes clustering problems, but it seems unintentional. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira