Guys, I created a simple implementation of IgnitePredicate<ClusterNode> based on user attributes. This is the most commonly used approach for node filters, but currently users are forced to create their own implementations each time. Would be useful to have it out of the box.
Can someone take a look at the implementation and tell me if there is anything wrong/missed? Ticket: https://issues.apache.org/jira/browse/IGNITE-4439 PR: https://github.com/apache/ignite/pull/1369 -Val
