On Fri, Feb 08, 2002 at 10:11:12AM -0600, Scott Hussey wrote: > To follow up on my previous question, I found a solution for us. It turns > out that if you have 127.0.0.1 in your /etc/hosts file, pvm binds to the > wrong address and then you get the "add <host>" -> terminate behavior. > Removing it and leaving only an entry for the intranet/internet address > for the machine allows it to work fine. I'm not sure what side-effects > this will have on the rest of the system. I guess we will find out. >
This was a known problem that used to be documented (just checked, isn't anymore). How did you end up with your hostname on the 127.0.0.1 address? I may have to put that discussion back into README.Debian. pvm forwards the raw IP number over the pvmd connections, using the hostname to look them up. It really needs a better method. It either needs more new upstream code to more intelligently find it's own IP address, or (the Debian workaround) the distribution/sysadmin needs to not put the hostname on the loopback interface in /etc/hosts. -Drake
