sshdfilter automatically blocks ssh brute force attacks by reading sshd log output in real time and adding iptables rules based on authentication failures. Block rules are created by logging on with an invalid user name, or wrongly guessing the password for an existing account. Block rules are removed after a week to maintain a small list of blocks. It also comes with a LogWatch filter.
Re: Actually...I'm REAL suspicious too Are you still using sshdfilter, and which version? I assume you aren't. In which case I'd suspect the sshd startup script is still running sshdfilter instead...
Re: Actually...I'm REAL suspicious too Hardly surprising, sshdfilter clears the SSHD chain of rules whenever it starts, to stop the rules going stale in the long term. You have created your problem...