[Hadoop-common-user] tasktracker takes long time to start?

I noticed a tasktracker (not a new node) starts to do hearbeating with job tracker after couple of minutes. What can cause this?

Here's log output when a task tracker is restarted after a clean shutdown of the cluster but it hangs for 9 minutes (network connection was good when the tt started):

...
2010-09-24 17:16:54,478 INFO org.apache.hadoop.ipc.Server: IPC Server handler 20 on 39197: starting
2010-09-24 17:16:54,486 INFO org.apache.hadoop.ipc.Server: IPC Server handler 21 on 39197: starting
2010-09-24 17:16:54,486 INFO org.apache.hadoop.mapred.TaskTracker: TaskTracker up at: localhost.localdomain/127.0.0.1:39197
2010-09-24 17:16:54,486 INFO org.apache.hadoop.mapred.TaskTracker: Starting tracker
tracker_hostname:localhost.localdomain/127.0.0.1:39197
2010-09-24 17:35:28,864 INFO org.apache.hadoop.mapred.TaskTracker: Starting thread: Map-events fetcher for all reduce tasks on tracker_hostname:localhost.localdomain/127.0.0.1:39197
...

Thanks!


Michael