Have you checked logs? What was the failure you saw? In an non-clean shutdown, the next time it starts it would try to recover the indicies. It takes time, and if the timeout value you set (for starting the service) is less than the time needed for the recovery, the server would have trouble starting. You can look at the logs (neo4j log and the message logs in your graph dir) to see if this is the case.
However, if it is the case, I'm still wondering how this can be acceptable in a production environment. For example, if the neo4j server failed and we have to start a new server with a snapshot of old data, chances are the old data snapshot is not going to be from a clean shut down. It sounds unacceptable to me to have hours of down time just because neo4j needs to recover the indicies. Any insights/best practices from experts? On Wed, Oct 26, 2011 at 10:58 AM, KanTube <[email protected]> wrote: > I encountered a odd error. I installed several windows updates (it was > Tuesday :) ) and then proceeded to reboot my servers (three in total). on > one server the Neo4j service did not restart. I then deleted the data > directory and the service started correctly. I then moved the data from > another server to this server and it started correctly again. Seems like > original data file became corrupted during the shutdown. I have the data > files if you think this is something you want to investigate (it is a very > small data set) > > setup win7 x64 (vm running in Fusion) > Neo4j 1.5M02 > > > > -- > View this message in context: > http://neo4j-community-discussions.438527.n3.nabble.com/1-5M02-failed-to-start-on-reboot-tp3455296p3455296.html > Sent from the Neo4j Community Discussions mailing list archive at > Nabble.com. > _______________________________________________ > Neo4j mailing list > [email protected] > https://lists.neo4j.org/mailman/listinfo/user > _______________________________________________ Neo4j mailing list [email protected] https://lists.neo4j.org/mailman/listinfo/user
