This project IS NO LONGER MAINTAINED. If you plan on using it, be ready to address any issues on your own and produce your own releases. Issues that ask others to do something will be ignored. Pull requests may be merged on a very random schedule.
This is a MongoDB-backed job store for the Quartz scheduler.
Artifacts are released to Bintray.
If you are using Maven, add the following repository
definition to your pom.xml
:
<repositories>
<repository>
<id>jcenter</id>
<url>https://jcenter.bintray.com/</url>
</repository>
</repositories>
With Gradle, add the following to your build.gradle
:
repositories {
maven {
url "https://jcenter.bintray.com/"
}
}
With Maven:
<dependency>
<groupId>com.novemberain</groupId>
<artifactId>quartz-mongodb</artifactId>
<version>2.2.0-rc2</version>
</dependency>
With Gradle:
compile "com.novemberain:quartz-mongodb:2.2.0-rc2"
Like most things in Quartz, this job store is configured
via a property file, quartz.properties
:
# Use the MongoDB store
org.quartz.jobStore.class=com.novemberain.quartz.mongodb.MongoDBJobStore
# MongoDB URI (optional if 'org.quartz.jobStore.addresses' is set)
org.quartz.jobStore.mongoUri=mongodb://localhost:27020
# comma separated list of mongodb hosts/replica set seeds (optional if 'org.quartz.jobStore.mongoUri' is set)
org.quartz.jobStore.addresses=host1,host2
# database name
org.quartz.jobStore.dbName=quartz
# Will be used to create collections like mycol_jobs, mycol_triggers, mycol_calendars, mycol_locks
org.quartz.jobStore.collectionPrefix=mycol
# thread count setting is ignored by the MongoDB store but Quartz requries it
org.quartz.threadPool.threadCount=1
When running in clustered mode, the store will periodically check in with the cluster. Should that operation fail, the store needs to decide what to do:
- Shut down
- Do nothing and optimistically proceed
Different strategies make sense in different scenarios. Pausing Quartz would be optimal but this job store currently doesn't have that option.
The org.quartz.jobStore.checkInErrorHandler.class
property controls the error handler
implementation.
To shut down the JVM (which is the default), add the following key to quartz.properties
org.quartz.jobStore.checkInErrorHandler.class=com.novemberain.quartz.mongodb.cluster.KamikazeErrorHandler
to ignore the failure:
org.quartz.jobStore.checkInErrorHandler.class=com.novemberain.quartz.mongodb.cluster.NoOpErrorHandler
If you use Quartzite or want your job classes to be available to Clojure code, use:
org.quartz.jobStore.class=com.novemberain.quartz.mongodb.DynamicMongoDBJobStore
(this assumes Clojure jar is on classpath).
By default you are allowed to pass any java.io.Serializable
objects inside JobDataMap
.
It will be serialized and stored as a base64
string.
If your JobDataMap
only contains simple types, it may be stored directly inside MongoDB to save some performance.
org.quartz.jobStore.jobDataAsBase64=false
To enable clustering set the following property:
# turn clustering on:
org.quartz.jobStore.isClustered=true
# Must be unique for each node or AUTO to use autogenerated:
org.quartz.scheduler.instanceId=AUTO
# org.quartz.scheduler.instanceId=node1
# The same cluster name on each node:
org.quartz.scheduler.instanceName=clusterName
Each node in a cluster must have the same properties, except instanceId. To setup other clusters use different collection prefix:
org.quartz.scheduler.collectionPrefix=yourCluster
Different time settings for cluster operations:
# Frequency (in milliseconds) at which this instance checks-in to cluster.
# Affects the rate of detecting failed instances.
# Defaults to 7500 ms.
org.quartz.jobStore.clusterCheckinInterval=10000
# Time in millis after which a trigger can be considered as expired.
# Defaults to 10 minutes:
org.quartz.jobStore.triggerTimeoutMillis=1200000
# Time in millis after which a job can be considered as expired.
# Defaults to 10 minutes:
org.quartz.jobStore.jobTimeoutMillis=1200000
# Time limit in millis after which a trigger should be treated as misfired.
# Defaults to 5000 ms.
org.quartz.jobStore.misfireThreshold=10000
# WriteConcern timeout in millis when writing in Replica Set.
# Defaults to 5000 ms.
org.quartz.jobStore.mongoOptionWriteConcernTimeoutMillis=10000
CI is hosted by Travis CI
(c) Michael S. Klishin, Alex Petrov, 2011-2020.
The project was originally started by MuleSoft. It supports all Quartz trigger types and tries to be as feature complete as possible.
MuleSoft developers did not respond to attempts to submit pull requests for several months. As more and more functionality was added and implementation code refactored, I decided to completely separate this fork form GitHub forks network because the project is now too different from the original one. All changes were made with respect to the Apache Public License 2.0.