RE: [Oscar-users] Orphan Nodes

[Bernard Li]

I cannot stress the importance of having backups of your headnode, since this is a single point of failure, you really want to make sure that you can somehow recover when you have hardware failures.   But that doesn't really help you right now though does it ;-)   Honestly, you are better off rebuilding the whole cluster.  You may be able to recover the SSH keys such that you don't need to enter a password to traverse between headnode and compute nodes, but your OSCAR installation will be fresh and will have no information about your nodes.  You also wouldn't have the image to push to your compute nodes.  Your package versions will be sqewed (unless you install the exact same version of OSCAR you had before).   So yeah, I think your best bet is to rebuild your cluster, unless you have a backup of the files of your headnode.   Try out the new version, you may like some of the new features like "Monitor Cluster Deployment" and better SATA support.   Cheers,   Bernard   From: [EMAIL PROTECTED] on behalf of Johnston Michael J Contr AFRL/DES Sent: Fri 25/11/2005 23:38 To: 'oscar-users@lists.sourceforge.net' Subject: [Oscar-users] Orphan Nodes I just posted in a previous listing that my hard drive on my head node blew up.  After reinstalling OSCAR on a new hard drive is it possible to bring all the existing nodes under the management of the rebuilt server without having to rebuild them?  I haven't tried yet, but I'm assuming that when I get the master node going again and I try to connect to any node it's going to want a password for each connection.  Am I wrong?  Thanks in advance!   Mike