Bob Proulx wrote: > I don't think using dhcp is correct because the network is already up > and assigned. But perhaps it is because that would allow the system > to migrate to a different address after the lease expires. I don't > know. I didn't do any testing of that configuration. (Maybe later.)
I tried a test using the "dhcp" configuration. Basically at boot the same behavior (not mounting nfs mounts) for the same reason because the network is already online and so isn't brought online and so doesn't run the if-up.d/* scripts. However it is much worse if you manually try to "ifup eth0" and bring the interface online manually. That causes a catastrophic failure! As far as I can see so far the first step in the process resets the network to an offline state and then brings starts the process to bring it back online. But since it is nfs diskless and requires the network the step bringing the network offline completely breaks the machine and it is incapable of pulling itself out of the problem. I can't really tell what is happening at that level because things become very unusable immediately but I think that is what is happening. Therefore I definitely do not recommend "dhcp". Setting that to "manual" avoids the potential breakage. At least in that case you can issue an "ifup eth0" okay. It runs the scripts okay. All is reasonable. Don't use a dhcp configuration for the main interface of an nfs diskless node. It is a killer. Bob
signature.asc
Description: Digital signature
