[ https://jira.jboss.org/jira/browse/JBTM-413?page=com.atlassian.jira.plugi... ] Mark Little commented on JBTM-413: ---------------------------------- Just to be less cryptic: currently during prepare phase the record is the last one to be triggered and if it succeeds we write the log record and move to phase 2. If it fails we abort. This optimises for the case where failures happen (we don't write the log unless we really have to - the LR could fail during commit). Obviously there's a situation where we could crash after the commit and before writing the log and in which case we're in a heuristic. If we shift things so that we do nothing during prepare for the LR (VoteCommit) and drive it to commit during the second phase then a) we write a log all the time, and b) we can still get a failure and end up in a heuristic. This second approach is what we used to do. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ https://jira.jboss.org/jira/browse/JBTM-413?page=com.atlassian.jira.plugi... ] Work on JBTM-413 started by Mark Little. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira