can't drop table

[shangan <sh...@corp.kaixin001.com>]

out of no disk space left, the namenode crash down. I try to restart the namenode but failed(either remove the edits or remove the last record in edits not work), finally I use another node as namenode and drop the old one,and remove data from all datanodes. That means I reconstruct a cluster, but I want to reuse the tables I created in hive,
when I do a "select * from tablename" there's an exception :
Failed with exception java.io.IOException:java.net.ConnectException: 
Call to vm125/10.11.2.125:9000 failed on connection exception: java.net.ConnectException: Connection refused 

I don't know why the hive still use the old namenode since I have use a new node as namenode, is there any information stored in the table when created ? if not, I think it might have errors as the directory corresponding to the table in the dfs has been removed , but it should not be the above errors. Can anybody explain it ?

btw, I can create new tables and operate on them, everything goes on well.


2010-08-31 



shangan