Red Hat Training

A Red Hat training course is available for Red Hat Fuse

Chapter 13. Beanstalk

Beanstalk component

Available in Camel 2.15
camel-beanstalk project provides a Camel component for job retrieval and post-processing of Beanstalk jobs.
You can find the detailed explanation of Beanstalk job life cycle at Beanstalk protocol.

Dependencies

Maven users need to add the following dependency to their pom.xml 
<dependency>
  <groupId>org.apache.camel</groupId>
  <artifactId>camel-beanstalk</artifactId>
  <version>${camel-version}</version>
</dependency>
where ${camel-version} must be replaced by the actual version of Camel (2.15.0 or higher).

URI format

beanstalk://[host[:port]][/tube][?options]
You may omit either port or both host and port: for the Beanstalk defaults to be used (“localhost” and 11300). If you omit tube, Beanstalk component will use the tube with name “default”.
When listening, you may probably want to watch for jobs from several tubes. Just separate them with plus sign, e.g. 
beanstalk://localhost:11300/tube1+tube2
Tube name will be URL decoded, so if your tube names include special characters like + or ?, you need to URL-encode them appropriately, or use the RAW syntax, see more details here.
By the way, you cannot specify several tubes when you are writing jobs into Beanstalk.

Common URI options

Name
Default value
Description
jobPriority 1000 Job priority. (0 is the highest, see Beanstalk protocol)
jobDelay 0 Job delay in seconds.
jobTimeToRun 60 Job time to run in seconds. (when 0, the beanstalkd daemon raises it to 1 automatically, see Beanstalk protocol)

Producer UIR options

Producer behaviour is affected by the command parameter which tells what to do with the job, it can be
Name
Default value
Description
command put
  • put means to put the job into Beanstalk. Job body is specified in the Camel message body. Job ID will be returned in beanstalk.jobId message header.
  • delete, release, touch or bury expect Job ID in the message header beanstalk.jobId. Result of the operation is returned in beanstalk.result message header
  • kick expects the number of jobs to kick in the message body and returns the number of jobs actually kicked out in the message header beanstalk.result.

Consumer UIR options

The consumer may delete the job immediately after reserving it or wait until Camel routes process it. While the first scenario is more like a “message queue”, the second is similar to “job queue”. This behavior is controlled by consumer.awaitJob parameter, which equals true by default (following Beanstalkd nature).
When synchronous, the consumer calls delete on successful job completion and calls bury on failure. You can choose which command to execute in the case of failure by specifying consumer.onFailure parameter in the URI. It can take values of bury, delete or release.
There is a boolean parameter consumer.useBlockIO which corresponds to the same parameter in JavaBeanstalkClient library. By default it is true.
Be careful when specifying release, as the failed job will immediately become available in the same tube and your consumer will try to acquire it again. You can release and specify jobDelay though.
Name
Default value
Description
onFailure bury Command to use when processing failed. You can choose among: bury, delete or release.
useBlockIO true Whether to use blockIO.
awaitJob true Whether to wait for job to complete before ack the job from beanstalk
The beanstalk consumer is a Scheduled Polling Consumer which means there is more options you can configure, such as how frequent the consumer should poll. For more details see Polling Consumer.

Consumer Headers

The consumer stores a number of job headers in the Exchange message:
Property
Type
Description
beanstalk.jobId long Job ID
beanstalk.tube string the name of the tube that contains this job
beanstalk.state string “ready” or “delayed” or “reserved” or “buried” (must be “reserved”)
beanstalk.priority long the priority value set
beanstalk.age int the time in seconds since the put command that created this job
beanstalk.time-left int the number of seconds left until the server puts this job into the ready queue
beanstalk.timeouts int the number of times this job has timed out during a reservation
beanstalk.releases int the number of times a client has released this job from a reservation
beanstalk.buries int the number of times this job has been buried
beanstalk.kicks int the number of times this job has been kicked

Examples

This Camel component lets you both request the jobs for processing and supply them to Beanstalkd daemon. Our simple demo routes may look like 
from("beanstalk:testTube").
   log("Processing job #${property.beanstalk.jobId} with body ${in.body}").
   process(new Processor() {
     @Override
     public void process(Exchange exchange) {
       // try to make integer value out of body
       exchange.getIn().setBody( Integer.valueOf(exchange.getIn().getBody(classOf[String])) );
     }
   }).
   log("Parsed job #${property.beanstalk.jobId} to body ${in.body}");
from("timer:dig?period=30seconds").
   setBody(constant(10)).log("Kick ${in.body} buried/delayed tasks").
   to("beanstalk:testTube?command=kick");
In the first route we are listening for new jobs in tube “testTube”. When they are arriving, we are trying to parse integer value from the message body. If done successful, we log it and this successful exchange completion makes Camel component to delete this job from Beanstalk automatically. Contrary, when we cannot parse the job data, the exchange failed and the Camel component buries it by default, so that it can be processed later or probably we are going to inspect failed jobs manually.
So the second route periodically requests Beanstalk to kick 10 jobs out of buried and/or delayed state to the normal queue.