The answer is automation (cfengine/puppet/bcfg2/chef/others). But those won't really help you unless you can keep them under strict control. You can easily obtain that in two steps:
- get a source code management repository running - i recommend subversion to start with
- integrate the scm with an automation tool of your choice
So move all your crons in /etc/cron.d/ and use a naming scheme that accomodates your systems. Second , put all /etc/cron.d/ directories in subversion and have local checkouts on each machine. You will be able to manage everything manually and have a history of your changes, be able to track any machines that do not have the latest crontabs and so forth. Then you can get an automation software to do all this for you.
If you really need something quick and dirty you should have a single file , let's say /etc/cron.d/all-jobs , deployed on on all machines and decide in that file if a cron is supposed to run or not based on the hostname.
Something like:
*/5 * * * * root [[ $HOSTNAME == 'myhostname' ]] && /usr/bin/job.sh
This is not very efficient but it should make your life easier until you get some automation going on.