Start of topic | Skip to actions
SSH configuration woes - problem and solution.This is related to twiki topic:CE overloaded and jobs not running. At the beginning of the week August 20 2007, I removed all pool accounts from charm-mgt and started afresh. The ATLAS VO wants separate pool accounts for production users and I found some time to do it. Since then we have been failing. This is what happened. At some point (don't remember when) I found some strangeness when ssh'ing around as a pool user. Sometimes it was challenged and other times it wasn't. So, to get around this I created ssh key pairs in each account and added the pub key to the authorised keys file in the account. When I re'gened them last week, I didn't recreate key pairs for the accounts. It turns out that hostbased Auth was not configured correctly so when the pbs script tried to stage in stuff via #PBS -W directives, it failed. What I did:
"short name, long name local interface,local ip address" ssh-rsa key.... "long name public interface, public ip address" ssh-rsa key...From the manpage of sshd: section SSH_KNOWN_HOSTS FILE FORMAT -- snip -- When performing host authentication, authentication is accepted if any matching line has the proper key. It is thus permissible (but not recommended) to have several lines or different host keys for the same names. This will inevitably happen when short forms of host names from different domains are put in the file.-- snip -- | |