In that case, the expectation from the broker is that disk io will result in an IOException and the default behavior in that case is to stop the broker. Look at: http://activemq.apache.org/configurable-ioexception-handling.html
On restart, once the partition is available, the broker will recover (replay the journal) from its last valid checkpoint. On 18 November 2011 15:29, Jeremy Deane <jde...@plymouthrock.com> wrote: > What are the implications to a live broker where the mount point containing > the KahaDB fails? And what is the recommended process to recover from such a > failure? > > -- > View this message in context: > http://activemq.2283324.n4.nabble.com/Live-Broker-and-KahaDB-on-NFS4-tp4083655p4083655.html > Sent from the ActiveMQ - User mailing list archive at Nabble.com. > -- http://fusesource.com http://blog.garytully.com
