0

I completed the code upgrade and removed all deprecated calls from my service code.

This is a Java Spring application and all services come up (as reported in the log files) when the Cluster (of 5 nodes) starts up.

However, when I try to get a serviceProxy to each service, all services get cancelled as a result of ServiceDeploymentTask attempting to do a redeploy of the services!! The redeploy fails and only cancels all of the services and fails to restart them. This can be demonstrated with both a thick and a thin client.

Why is Ignite trying to redeploy the services? (and why don't the services restart?)

Is there something I'm missing from the move to Ignite 2.10???

Finally, why does a Java Thin Client create a NODE_JOIN event?

Thanks in advance. Greg

Greg Sylvain
  • 1
  • Is it possible to share the logs? Java thin client won't trigger EVT_NODE_JOINED because a thin node can't be a part of the topology, something wrong is happening here. – Alexandr Shapkin Jan 02 '22 at 21:19
  • Sorry, I can't scrub the logs and get them out. But, that is what I thought (thin clients should not generate topology changes). What would cause a redeploy to be called/queued, any idea? – Greg Sylvain Jan 03 '22 at 22:03
  • @GregSylvain can you share a reproducer for your issue? Afaik thin client appearance should not emit NODE_JOIN event. – solveMe Jan 10 '22 at 06:28
  • Please provide enough code so others can better understand or reproduce the problem. – Community Jan 11 '22 at 12:25
  • Igniters, I found my issue with the thin client kicking off a NODE_JOIN host. (In my refactoring of a thick client, I left the Spring bean connection outside of the _if thin client_ block.) Sorry for the erroneous scenario. – Greg Sylvain Jan 15 '22 at 15:19

0 Answers0