(taken from the Planet-Lab documentation)
For a variety of reasons, including hardware failure and security compromise, nodes come and go. If a node in your slice is down permanently or temporarily, you will not be able to login to it. If the same node returns to service and you have not removed it from your slice, you will be able to access it when it becomes available again. Be aware that any node may be re-installed or even replaced at any time and that, consequently, data should not be permanently stored on any node.
If you are not able to login to a particular node that you are sure has been assigned to your slice, verify that you attempted to login with your slice name as the user name. Add -v to the ssh command line to print more information about the connection attempt.
ssh -v -l princeton_test1 -i ~/.ssh/id_rsa planetlab-1.cs.princeton.edu
Another common problem is mismatched host keys. The SSH host keys of PlanetLab nodes occasionally change. If you have a stale entry in your host key cache (by default, located at ~/.ssh/known_hosts), replace it with the current entry for the host which you may obtain from the Known Hosts page. You may merge this file directly into your ~/.ssh/known_hosts file if you wish.
If you are still unable to login, consult at least one of the community-supported Status Tools before asking PlanetLab Support for help. The node may simply be down for an extended period of time for known reasons. When contacting PlanetLab Support about a failed login attempt, include your slice name, the name of the node, and the verbose output of your ssh attempt.