Hi, You just need to specify cluster group:
ignite.compute(ignite.cluster().forServers()).run(new IgniteRunnable() { @Override public void run() { System.out.println("This will be processed on server nodes only"); } }); There a various of possible group filters, for example you may run on data nodes, that participate in cache to avoid data transferring from affinity nodes: ClusterGroup cacheGroup = ignite.cluster().forCacheNodes("myCache"); If you use PARTITIONED cache some cache queries in compute jobs may also get entries from remote node. You may change cache to REPLICATED or use ignite.compute().affinityRun(). In this case compute job will be invoked on node that has this entry. -Dmitry. -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Compute-calls-happening-on-client-node-tp12011p12020.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.