Welcome to WebmasterWorld Guest from 18.104.22.168
The first thing to do is to block this particular 'bot. Then investigate how it is that it thought it was allowed to spider parts of your site that you don't want spidered -- Is there a problem or a logical inconsistency in your use of robots.txt and on-page meta-robots tags, for example? Or does it simply not obey your valid robots.txt and on-page robots-control directives?
For legitimate but unwanted 'bots that always identify themselves, Robots.txt and User-Agent-string -based access controls are the 'least expensive' in terms of list size and server performance impact. Next would be large IP address-range blocks (but be careful, as the correlation of ranges to countries, regions, or companies is often very poor), and finally, smaller specific IP address ranges for specific robots.
Another avenue to pursue is that of whitelisting. Rather than try to deny everything that's 'bad,' consider allowing only those accesses that you deem to be 'good'. Careful analysis of incoming requests, looking at *all* of the HTTP headers sent with each request (many of which do not show up in standard server logs) can be quite telling. For example, some Googlebot requests are fake, and don't come from Google at all. But this isn't obvious until you check some of the additional HTTP headers not usually logged by servers.
Also take a look at the bad-bot scripts posted here by Key_Master, xlcus, and AlexK over in the PHP and PERL forum libraries. These implement behavior-based access control methods, and provide yet another 'angle' on the access control problem.
The subject of access control including robot control and the more-general user-agent control is quite wide and very deep. So it's a good idea to address your immediate problem and then spend a year or so developing what you feel is an appropriate access-control policy for your site. After 'paying attention' to this problem for an extended period of time, you'll find that what you want and need to do becomes quite a bit clearer. As a result, you'll waste far less time implementing great solutions to the wrong problems.
I guess this is a very long way to say that there is no one-size-fits-all solution, and it wouldn't be wise to just try to copy-and-paste a solution off "some Web site on the internet."