order of operations when adding a node to an existing cluster
Whenever I have a running cluster and add new nodes, I notice after the
node is added to SGE, the users and scratch space is set up. Occasionally
a job fails on a node because of this order. Either the user who submitted
the job isn't set up in /etc/password yet or the scratch space hasn't been
created yet. Why isn't a node fully configured before its added to SGE?
That would seem like the better order.
Ryan
Received on Thu Apr 03 2014 - 15:09:08 EDT
This archive was generated by
hypermail 2.3.0.