We are having a very frustrating problem with SSH on 3 11i servers

[Christina Martin]

We’ve upgraded 8 so far. The first 5 had absolutely NO problems at all. The load for the last 3 servers was identical (we used the exact SAME ignite tape – only changing the network information which has been verified over and over again. (again, same procedures, no problems with the first 5 servers)

When we loaded the last 3 servers we started getting an error

Invalid SSH exchange identification!
Check your host profile protocol type.


The sshd_config file on all 8 servers is identical.
When you do a ps –ef | grep ssh on the last 3 servers
You get numerous sshd processes that have spawned but aren’t really there (zombies).
When you try to kill the process, it doesn’t exist. We’ve stopped the ssh daemon (/sbin/init.d/secsh stop) most of the processes go away. When we restart the daemon (/sbin/init.d/secsh start) the ALL come back.

This problem is intermittent. Some can log on immediately, but when they log off it takes 5, 6, 10 tries before the system lets us in. WE’ve done the ssh –vvv –l and don’t get anymore information than the error that I’ve noted above.

Any information that you can provide would be greatly appreciated. We have one more server to upgrade and have to have this worked out or our site won’t let the upgrade continue...

[Christina Martin]

this ended up being an application problem