Question: Issue with Remote Event listener on Cluster Nodes

Hi

We are using ignite services to register Remote Events listener to listen to cache updates.

The behavior we are currently seeing with above design is when we deploy the service on one node cluster(1 Node only) the service functions as expected

but when we have more than one node in cluster and the service is deployed to one or more we are getting below ERROR

[sys-#51][GridDeploymentPerVersionStore] Failed to get resource from node [nodeId=d88e4292-49f7-408d-b410-b26f98df3335,
clsLdrId=28c72479191-d88e4292-49f7-408d-b410-b26f98df3335, resName=com/example/ignite/CacheRemoteEventService$$Lambda/0x00007f3b48717878.class,
classLoadersHierarchy=org.apache.ignite.internal.managers.deployment.GridDeploymentClassLoader->jdk.internal.loader.ClassLoaders$AppClassLoader->
jdk.internal.loader.ClassLoaders$PlatformClassLoader, msg=Requested resource not found (ignoring locally)
[originatingNodeId=f5b7438a-ebec-4d96-a55e-2e9fa1c6edb8, resourceName=com/example/ignite/CacheRemoteEventService$$Lambda/0x00007f3b48717878.class,
classLoaderId=28c72479191-d88e4292-49f7-408d-b410-b26f98df3335]]


Below is the sample configuration for Services


We are converting the CacheRemoteEventService implementation into jar and uploading it to libs folder of every Ignite NODE.

Tried multiple scenarios to deploy and execute the service but all of them shows the same ERROR.

Scenarios tried which fails

1. Uploading service library on all nodes and having service config on only one node that way service is only deployed in one node.
2. Uploading service library on all nodes and having same service config on all available nodes that way service is deployed on each node in cluster.

Could you please suggest if we are doing something wrong with respect to the design or configuration.

Start by reviewing the CacheRemoteEventService deployment configuration and the GridDeploymentPerVersionStore error for the missing lambda resource. Compare the one-node and multi-node service setups described in the issue, then reproduce the failure with the service library and configuration on multiple nodes. Done means the remote event listener deploys without the reported resource-loading error, or the required configuration is documented.

Written by the indexing model from the issue text.