The client is in this case an automated test, which issues REST requests to the application (which hosts the grid). The fact that the client joins the ignite cluster as a client with a sole purpose of pre-populating some caches is irrelevant to my question. What is happening here is that queries are being broadcasted to the newly started node prematurely as soon as it comes online. A seeming workaround for this would be to hold off all queries from the time when the new node is discovered to the time when its cache setup is completed. I wonder if you could suggest how to solve this problem, perhaps by subscribing to a series of events?
Thank you very much! -- View this message in context: http://apache-ignite-users.70518.x6.nabble.com/Exception-during-node-initialization-tp1518p1522.html Sent from the Apache Ignite Users mailing list archive at Nabble.com.
