Thanks to Ed and Nitin. Even after setting trash interval to zero, I get the following error message. I shall try the brute force "dfs -rmr skipTrash.." now.
What does this message mean?: FAILED: Error in metadata: MetaException(message:Got exception: java.io.IOException Failed to move to trash: hdfs:// xxxxxx.xxx.com/hive/max_sum_tbl) FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.DDLTask I saw the logs as well and the same error message was there. I did not see any additional information as to why it "failed to move to trash". I am assuming that it occurs if the user directory doesn't have enough space ...please correct me. Thanks, PD. On Mon, Dec 24, 2012 at 5:37 AM, Edward Capriolo <edlinuxg...@gmail.com>wrote: > Well in hive you can set hadoop variables so you can use the set command > to explicitly disable trash on the client > > set x=y > > If that does not work it is possible to run dfs commands from hive. Just > do your normal hadoop dfs command without specifying hadoop. > > hive > dfs -rmr /user/hive/warehouse > > > On Mon, Dec 24, 2012 at 2:46 AM, Periya.Data <periya.d...@gmail.com>wrote: > >> Hi, >> Is there a way to drop a table such that the contents do not go to the >> .Trash dir? I have limited diskQuota in hadoop and am running large Hive >> jobs in sequence. I would like to drop tables as and when they become >> unnecessary and also they must not end up in .Trash..as they occupy lot of >> space. >> >> Before I begin my Hive job, I do clean out (hadoop fs -rmr -skipTrash >> ...). I would like to know if there something like this to add in my hql >> file. >> >> Thanks, >> PD. >> >> >> >
