Bad /bin/init recovery?

Is re-imaging the root directory the only way to recover from a bad init? I got my D.net client running all the time, and thats part of the problem. I forgot to background the buffer checker, so the unit boots and runs the client and keeps looping there. Can the system be booted in a way to ignore the init once and dump to a shell? Or can the init be the only thing wiped by using the pump program somehow? I'd rather not lose the work I didn't backup. (I'll be making a backup script once I get back to a shell prompt to prevent this from hindering me in the future.)

Thanks

My empeg site is:http://24.236.3.131/empeg/

The only way I can think of is to compile your own kernel with "init=/bin/sh" as part of the kernel command line (set using make config).


--
Mike Crowe
I may not be speaking on behalf of empeg above :-)

I eventually gave up and reloaded. Took me a while to remember all of what I had changed and added, but now that I know about /dev/hda2 being usable, it won't happen again. Right now, I have a backup script that dumps /usr /var /etc and /bin to that partition, and a restore script ready to Z-Modem back over. Had to use it once again last night, working with init at 2am is not the best idea in the workd :-)

My empeg site is:http://24.236.3.131/empeg/