RE: [beagleboard] Problem: A start job is running for LSB ... [fixed]

OK. So I finally got this all working (with the Debian Jessie RFS no less) by exporting a file system for each target board. So overall directory structure is something like this:

/home/bone/tftp <-- vmlinuzXXX and /dtbs/XXXX for TFTP boot
/home/bone/<mac_suffix> <- 1st NFS root file system export
/home/bone/<mac_suffix> <- 2nd NFS root file system export

I think I have an idea why the failures with the original shared layout.

Examining the Wireshark traffic I eventually noticed the following: when the blocking DHCP request packet was sent it was always requesting the same IP address (.111) and *not* the IP address used to boot and mount (.114 for example). For whatever reason the router never responded to this request (even if the address was unallocated).

Could this be down to the file system having some cached state?

Also, FWIW, I did run up a Debian Wheezy RFS via NFS. It asked me for a runlevel and then printed 'unabl' after logon credentials were entered. Decided life was too short at this point and abandoned experiment.

Thanks again for all the input. Much appreciated.


/home/bone/tftp ā† vmlinuzXXX and /dtbs/XXXX for TFTP boot
/home/bone/<mac_suffix> ā† 1st NFS root file system export
/home/bone/<mac_suffix> ā† 2nd NFS root file system export

Only thing I can think of here is that when a board mounts, and takes over a rootfs, it somehow becomes locked elusively for that board. This kind of makes sense but, not really. Iā€™d have to do some reading on the subject to remember how all this works, and quite honestly. I do not want to, as I have other things to worry about. :confused: