Why does the oplog slurper start with latest oplog ts?

[dharshanr]

Should it not start with the last processed timestamp instead of the latest timestamp?

            Shard shard = config.getShards().get(0);
            Thread tailerThread = EsExecutors.daemonThreadFactory(
                    settings.globalSettings(), "mongodb_river_slurper_" + shard.getName() + ":" + definition.getIndexName()
                ).newThread(new OplogSlurper(shard.getLatestOplogTimestamp(), mongoClusterClient, mongoClusterClient, definition, context, esClient));
            tailerThreads.add(tailerThread);                   

Here is the scenario

1. Configure oplog tailing
2. Stop the elastic server (which operations are still going on)
3. Restart the server

The sync starts from the latest operation and not the last operation which means the intermediate operations are lost.

[benmccann]

This should be fixed now in the next release. Thank you for reporting it! It was very helpful