Watch out, if you're Linux or something similar, cron probably runs tmpwatch periodically which might be deleting old files from /tmp.
Regards, Albert On Feb 11, 2008 8:43 PM, Raghu Angadi <[EMAIL PROTECTED]> wrote: > > Did you find /tmp/hadoop-root/dfs/name earlier? Are you sure the config > hasn't changed before the last restart? Hadoop itself does not remove > these directories for any reason. > > Raghu. > > > Ben Kucinich wrote: > > Many times while starting hadoop, bin/start-all.sh, I get this error > > in logs/hadoop-root-namenode-master.log
