[StarCluster] weird errors

François-Michel L'Heureux fmlheureux at datacratic.com
Tue Sep 3 09:22:01 EDT 2013


Hi Kai

About your 20 nodes cluster, was node001 up all the time or did it
stop/restart at any point? If it did, then the node references need to be
cleaned and recreated as the identification does indeed change.

The way to solve this is simple: run "ssh-keygen -f
"/home/kli/.ssh/known_hosts" -R node001" as it mentions.

Was a StarCluster load balancer running? I know this error happens from
times to times but I haven't spent time over it.

Mich
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.mit.edu/pipermail/starcluster/attachments/20130903/95472427/attachment.htm


More information about the StarCluster mailing list