1) Expected exception stack trace is logged at warning level. This causes the monitoring system(EKG) to pickup these stack traces. Do not log these stack trace unless debug is enabled or if the exception has inner exception 2) When a node is unavailable in the resource pool map, ignore the cleanup. This can happen because the resource pool map used by the failure detector and the other requests are different.
edited: PropsTest.java now is simpler and does, pretty much the same thing
next: more tests
1) Refactored the code and introduced two types of comparison version based and hash of the value based 2) The code was dependent on node 0 for boot strapping. fixed. 3) Enable the test testDetermineConsistencyHashValue and fixed it. 4) Now the consistency check tool takes a new option --comparison-type which defaults to version and can be forced to hash based comparison.
…y file to depend on tusk-0.0.2. tusk-0.0.1 jar was broken and needed to be updated.