The way I handle this is I greylist any email that comes in if:
1) it has no reverse dns
2) or the reverse dns matches certain patterns that indicate dynamic ip's

The rest comes in undelayed and is exposed to further checks. Note that it's ludicrous to require forward and reverse dns to resolve to the same thing. In today's age of hosting multiple domains in one server/cluster, it's just not reasonable to have PTR records defined that way.

You're blaming the wrong entity. If you're concerned with this, you should be complaining to your ISP _whom_you_pay_ that they use Spamhaus. You have control of your service, go buy it from someone who doesn't use Spamhaus. Spamhaus isn't screwing with your Inbox, your ISP _whom_you_pay_ is screwing with your Inbox by their choice to use Spamhaus.

Don't get me wrong, I think Spamhaus is one of the best things since sliced bread. Why does your ISP _choose_ to use Spamhaus? Because the extra cost and resources involved with NOT using Spamhaus would impact their bottom line and they would have to charge you more.

Before all the botnet takedowns, RBL's used to account for blocking about 80-85% of inbound connections. Now it's down to less than 50%.
$ emailstats
Webmail System Statistics for 2011-10-12

  TotalIncoming: 187662
                      RBL: 100601
                  Spams: 19439
              Viruses: 192
            Accepted: 67430
LocalDelivered: 53243
          Forwarded: 14187
      PercentGood: 35.9316

Chances are you didn't do the two main things that Linux requires for changing hardware out from underneath it:
1. Regenerate the modules needed at boot to access the harddrive(s). Cent/RH/Mandriva all have grub load an initrd with modules that the kernel needs to access the controller(s) and ...
2. Change the module needed for network hardware.

Other possible issues are network device mac address change (Cent/RH tend to hardcode the MAC address into the interface config file), hostname changes, network interface reordering, adjust boot parameters to be friendly inside vmware, etc. These are small though compared to the two main ones above.