The issue appears to persist in the new version. I have ignore_older and close_older set to 1 hour and I am still running into inode re-use which is causing filebeat to think that a rename is occurring. This issue appears to be based on filebeat thinking that I am updating a previously used inode. The same types of log entries I posted before are present in the logs and I still have data gaps in kibana when this is occurring. Here is an example of what is in my log:
2016-04-01T20:20:04Z INFO Registry file updated. 9399 states written.
2016-04-01T20:20:04Z INFO Old file with new name found: /mydata/logs-20160228.2300.data is no /mydata/logs-20160401.2010.data
2016-04-01T20:20:04Z INFO Detected rename of a previously harvested file: /mydata/logs-20160228.2300.data -> /mydata/logs-20160401.2010.data
2016-04-01T20:20:04Z INFO Harvester started for file: /mydata/logs-20160401.2010.data
2016-04-01T20:20:04Z INFO Registry file updated. 9400 states written.