Description of problem: If there are timeout problems during replication the cache can become inconsistent for different nodes. But there is no information which entries are affected. There should be a warning for the inconsistency and an information which key is affected to be able to do a cleanup.
For the different commands it might not possible to provide that information. Also it might not possible to print a useful information if the key is a binary value or did not contain a useful toString() method for custom classes.
Some log messages already print the key (such as lock timeouts), so that part shouldn't be an issue.
PR: https://github.com/infinispan/jdg/pull/775
Error log records now contains affected keys, log record looks like this (in this case the key was byte array so strange key name is expected:-) Error executing command PutKeyValueCommand, writing keys [[B@232f4a48]: org.infinispan.util.concurrent.TimeoutException: Replication timeout for server1/clustered
This product has been discontinued or is no longer tracked in Red Hat Bugzilla.