Lesson 11: Advanced (Enterprise) Features
Clustering currently works with the JDBC-Jobstore (JobStoreTX or JobStoreCMT) and the TerracottaJobStore. Features include load-balancing and job fail-over (if the JobDetail’s “request recovery” flag is set to true).
Clustering With JobStoreTX or JobStoreCMT
Enable clustering by setting the “org.quartz.jobStore.isClustered” property to “true”. Each instance in the cluster should use the same copy of the quartz.properties file. Exceptions of this would be to use properties files that are identical, with the following allowable exceptions: Different thread pool size, and different value for the “org.quartz.scheduler.instanceId” property. Each node in the cluster MUST have a unique instanceId, which is easily done (without needing different properties files) by placing “AUTO” as the value of this property.
Never run clustering on separate machines, unless their clocks are synchronized using some form of time-sync service (daemon) that runs very regularly (the clocks must be within a second of each other). See http://www.boulder.nist.gov/timefreq/service/its.htm if you are unfamiliar with how to do this.
Never fire-up a non-clustered instance against the same set of tables that any other instance is running against. You may get serious data corruption, and will definitely experience erratic behavior.
Only one node will fire the job for each firing. What I mean by that is, if the job has a repeating trigger that tells it to fire every 10 seconds, then at 12:00:00 exactly one node will run the job, and at 12:00:10 exactly one node will run the job, etc. It won’t necessarily be the same node each time - it will more or less be random which node runs it. The load balancing mechanism is near-random for busy schedulers (lots of triggers) but favors the same node that just was just active for non-busy (e.g. one or two triggers) schedulers.
Clustering With TerracottaJobStore
Simply configure the scheduler to use TerracottaJobStore (covered in Lesson 9: JobStores), and your scheduler will be all set for clustering.
You may also want to consider implications of how you setup your Terracotta server, particularly configuration options that turn on features such as storing data on disk, utilization of fsync, and running an array of Terracotta servers for HA.
More information about this JobStore and Terracotta can be found at http://www.terracotta.org/quartz
As explained in Lesson 9: JobStores, JobStoreCMT allows Quartz scheduling operations to be performed within larger JTA transactions.
Jobs can also execute within a JTA transaction (UserTransaction) by setting the “org.quartz.scheduler.wrapJobExecutionInUserTransaction” property to “true”. With this option set, a a JTA transaction will begin() just before the Job’s execute method is called, and commit() just after the call to execute terminates.
Aside from Quartz automatically wrapping Job executions in JTA transactions, calls you make on the Scheduler interface also participate in transactions when using JobStoreCMT. Just make sure you’ve started a transaction before calling a method on the scheduler. You can do this either directly, through the use of UserTransaction, or by putting your code that uses the scheduler within a SessionBean that uses container managed transactions.