Hi, We are using Hadoop 1.0.3 on Ubuntu 12.04.2 LTS. Hadoop servers include 1 NN/JT, 1 SNN/DN & several DNs.
>From time to time, some of the servers just hanged, cannot be pinged, screen blackened out, not responding to keyboard input and lost connection with the NN. Lately, one DN was hanged even when there is no job to run. Specifically, the unresponsive happens not on all machines. It usually happens on several specific DNs. How to tackle this problem? Does it leave a trace when the system crashes/hangs? Any help would be greatly appreciated.
