nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

As we transition to the new Hyland Community, some post links may be temporarily disrupted as we finalize redirects. We apologize for the inconvenience and are working to resolve this as quickly as possible. Details.

Cluster aware scheduled jobs

nikes

Hello,

We have setup a cluster of 2 nodes (active/active) and would like to make our custom scheduled jobs cluster aware. Means
Job will be configured on both the node, but should get executed on one node only and not on both.

I am exploring JobLockService for this. any one tried before and would like to share example code?

Thanks

Find more posts tagged with

Comments

mrogers

The JobLockService is the way to do this.

Basically the jobs can execute on any node in the cluster but only one at a time. It allows you to take a system wide lock with a timeout (so if your system is shut down or stops the lock will expire.) And in addition you can extend the lock as your job progresses, you can either do that explicitly or you can set up a callback that calls you back to say is your job still alive.

Look at the alfresco source code for examples of its use.

nikes

Thanks mrogers

I am able to run jobs using JobLockService now.

For anyone looking for an example check these code

http://svn.alfresco.com/repos/alfresco-open-mirror/alfresco/HEAD/root/projects/repository/source/java/org/alfresco/repo/search/impl/solr/SolrBackupClient.java

http://svn.alfresco.com/repos/alfresco-open-mirror/alfresco/HEAD/root/projects/repository/source/java/org/alfresco/repo/content/cleanup/ContentStoreCleaner.java

patil

Another way of doing it is by extending the AbstractScheduledLockedJob class and overriding
   public void executeJob(JobExecutionContext context)
         throws JobExecutionException {} method.

The parent class manages locking using the method

/**
     * Lazily update the job lock
     */
    private void refreshLock()
    {
        Pair<Long, String> lockPair = lockThreadLocal.get();
        if (lockPair == null)
        {
            String lockToken = jobLockService.getLock(lockQName, LOCK_TTL);
            Long lastLock = new Long(System.currentTimeMillis());
            // We have not locked before
            lockPair = new Pair<Long, String>(lastLock, lockToken);
            lockThreadLocal.set(lockPair);
        }
        else
        {
            long now = System.currentTimeMillis();
            long lastLock = lockPair.getFirst().longValue();
            String lockToken = lockPair.getSecond();
            // Only refresh the lock if we are past a threshold
            if (now - lastLock > (long) (LOCK_TTL / 2L))
            {
                jobLockService.refreshLock(lockToken, lockQName, LOCK_TTL);
                lastLock = System.currentTimeMillis();
                lockPair = new Pair<Long, String>(lastLock, lockToken);
                lockThreadLocal.set(lockPair);
            }
        }
    }

Thanks,
Patil

mkopacz

Hi, I have a cluster consisting of two machines with Alfresco 5.0 Enterprise installed and would like to create a job that gets executed on only one of these machines at a time. I have followed this article http://docs.alfresco.com/5.1/references/dev-extension-points-scheduled-jobs.html, but without success - my job executes simultaneously on both nodes. Do you have any suggestions?