EJB cluster timer
Multiple application server instances can be arranged in a cluster that implements a distributed EJB timer system. A high availability database can store EJB timer info. The EJB timer jobs can be distributed among the application server instances of the cluster that implement the distributed EJB timer system. In case of a failure of an application server instance, the EJB timer info can be used to reassign the EJB timer jobs associated with the failed application server instance.
1. A computer implemented system comprising:
one or more computers, each including a microprocessor;
an application server cluster executing on the one or more computers and including a plurality of application server instances, wherein the cluster implementing a distributed Enterprise Java Bean (EJB) timer system and providing an EJB timer application program interface (API) receives requests for timer jobs to be processed by the cluster, each timer job created by implementing a class and including an EJB timer configured to trigger an execution of an EJB deployed on the cluster when the corresponding EJB timer expires, and wherein information for the EJB timers is serialized and stored in a database which additionally stores classloader information of the classes implemented by the timer jobs; and
a timer master that executes as a singleton service on one of the plurality of application server instances, each of which application server instances requests timer jobs from the timer master, wherein the timer master is configured to
retrieve the information for the EJB timers from the database and deserialize the information,
use the deserialized information to distribute the timer jobs among the application server instances,
perform, for each timer job to be distributed, the steps of
identifying, from the plurality of application server instances, an application server instance configured to execute the particular EJB associated with that timer job,
retrieving, from the database, classloader information of the class implemented by that timer job,
placing the classloader information of the implemented class on the identified application server instance to load the class thereon, wherein the placing utilizes the retrieved classloader information to load the corresponding EJB timer, and
upon failure of the identified application server instance, use the information for the loaded EJB timer on the identified application server instance to reassign the distributed timer job from the identified application server instance, to another of the application server instances.
2. The computer implemented system of claim 1 , wherein the application server instance that hosts the timer master maintains a lease for the timer master in a lease table in a database.
3. The computer implemented system of claim 2 , wherein upon failure of the application server instance currently hosting the timer master, another application server instance in the cluster is reassigned to host the timer master.
4. The computer implemented system of claim 3 , wherein the assignment of another application server instance to host the timer master functionality is performed by a migration master.
5. The computer implemented system of claim 1 , wherein the timer master schedules jobs in addition to the timer jobs.
6. The computer implemented system of claim 1 , wherein each of the application server instances is configured to view all of the scheduled timer jobs via the database.
7. The computer implemented system of claim 1 , wherein the cluster contains an additional application server instance in addition to the plurality of application server instances.
8. A non-transitory computer readable storage medium including instructions stored thereon, which when read and executed by one or more computers cause the one or more computers to perform the steps comprising:
configuring an application server cluster executing on the one or more computers and including a plurality of application server instances, wherein the cluster implementing a distributed Enterprise Java Bean (EJB) timer system and providing an EJB timer application program interface (API) receives requests for timer jobs to be processed by the cluster, each timer job created by implementing a class and including an EJB timer configured to trigger an execution of an EJB deployed on the cluster when the corresponding EJB timer expires, and wherein information for the EJB timers is serialized and stored in a database which additionally stores classloader information of the classes implemented by the timer jobs; and
configuring a timer master that executes as a singleton service on one of the plurality of application server instances, each of which application server instances requests timer jobs from the timer master, wherein the timer master is configured to
retrieve the information for the EJB timers from the database and deserialize the information,
use the deserialized information to distribute the timer jobs among the application server instances,
perform, for each timer job to be distributed, the steps of
identifying, from the plurality of application server instances, an application server instance configured to execute the particular EJB associated with that timer job,
retrieving, from the database, classloader information of the class implemented by that timer job,
placing the classloader information of the implemented class on the identified application server instance to load the class thereon, wherein the placing utilizes the retrieved classloader information to load the corresponding EJB timer, and
upon failure of the identified application server instance, use the information for the loaded EJB timer on the identified application server instance to reassign the distributed timer job from the identified application server instance, to another of the application server instances.
9. A method of providing an EJB cluster timer, comprising:
providing an application server cluster executing on the one or more computers and including a plurality of application server instances, wherein the cluster implementing a distributed Enterprise Java Bean (EJB) timer system and providing an EJB timer application program interface (API) receives requests for timer jobs to be processed by the cluster, each timer job created by implementing a class and including an EJB timer configured to trigger an execution of an EJB deployed on the cluster when the corresponding EJB timer expires, and wherein information for the EJB timers is serialized and stored in a database which additionally stores classloader information of the classes implemented by the timer jobs; and
providing a timer master that executes as a singleton service on one of the plurality of application server instances, each of which application server instances requests timer jobs from the timer master, wherein the timer master is configured to
retrieve the information for the EJB timers from the database and deserialize the information,
use the deserialized information to distribute the timer jobs among the application server instances,
perform, for each timer job to be distributed, the steps of
identifying, from the plurality of application server instances, an application server instance configured to execute the particular EJB associated with that timer job,
retrieving, from the database, classloader information of the class implemented by that timer job,
placing the classloader information of the implemented class on the identified application server instance to load the class thereon, wherein the placing utilizes the retrieved classloader information to load the corresponding EJB timer, and
upon failure of the identified application server instance, use the information for the loaded EJB timer on the identified application server instance to reassign the distributed timer job from the identified application server instance, to another of the application server instances.
10. The method of claim 9 , wherein the application server instance that hosts the timer master maintains a lease for the timer master in a lease table in a database.
11. The method of claim 10 , wherein upon failure of the application server instance currently hosting the timer master, another application server instance in the cluster is reassigned to host the timer master.
12. The method of claim 11 , wherein the assignment of another application server instance to host the timer master functionality is performed by a migration master.
13. The method of claim 9 , wherein the timer master schedules jobs in addition to the timer jobs.
14. The method of claim 9 , wherein each of the application server instances is configured to view all of the scheduled timer jobs via the database.
15. The method of claim 9 , wherein the cluster contains an application server instance in addition to the plurality of application server instances.
16. The computer implemented system of claim 1 , wherein the cluster contains an additional application server instance in addition to the plurality of application server instances.
17. The computer implemented system of claim 1 , wherein the system enables scheduled jobs to be submitted to be run within the cluster.
18. The method of claim 9 , wherein scheduled jobs can be submitted to be run within the cluster.