realtime node not removing sements after handoff

Hi,
My cluster was running fine and as expected.
The disk got full for realtime node and historical node which I resolved my moving whatever indexcache and basepersist data they created to other disk and giving that path in the respective config files.
The strange behavior I am seeing now is realtime node is not deleting segment even after they are loaded in historical node. There is a overlap of the data in these two nodes.If I query realtime node I get the following time range:
[ {
“timestamp” : “2015-01-01T09:40:00.000Z”,
“result” : {
“minTime” : “2015-01-01T09:40:00.000Z”,
“maxTime” : “2015-01-01T22:10:13.000Z”
}
} ]

``

If I query my historical node I get following time range:
[ {
“timestamp” : “2015-01-01T00:00:00.000Z”,
“result” : {
“minTime” : “2015-01-01T00:00:00.000Z”,
“maxTime” : “2015-01-01T19:14:59.000Z”
}
} ]

``

My window period is 5 min and intermediate persist period is 10m.

where could be the problem. Is realtime not able to understand that data has been loaded in the historical node. which agent sends this message to realtime node?
This is really problematic as I am not getting the correct performance. Please advice

while querying the broker node it gives me full time range:
[ {
“timestamp” : “2015-01-01T00:00:00.000Z”,
“result” : {
“minTime” : “2015-01-01T00:00:00.000Z”,
“maxTime” : “2015-01-01T22:10:13.000Z”
}
} ]

``

How long have you waited for the handoff to occur?

Please see:

http://druid.io/docs/0.8.3/ingestion/faq.html for handoff debugging