The solution is to run mount -a -t nfs in /etc/rc.local. Depending on your aesthetics, this might be considered a work-around.
More details below. On Sun, 2012-12-09 at 01:19 -0700, Bob Proulx wrote: > Ross Boylan wrote: > > I have a theory that the mounts are supposed to happen when the network > > device comes up; the regular network up routines are not triggered to > > avoid screwing up the root fs. /etc/network/interfaces has > > > > # The primary network interface > > # do not bring up interface twice--PXE already did it > > #allow-hotplug eth0 > > iface eth0 inet dhcp > > > > Is my theory correct? > > Yes. You are correct. Thanks for confirming. > Mostly. I fired up my nfs diskless client and > debugged through the startup process. The eth0 network interface is > already up at the start and therefore isn't ever brought up. The nfs > mounting is normally done in the /etc/network/if-up.d/mountnfs script. > But since the interface isn't brought up that script is never run. It > is run for the lo loopback device but never run for the eth0 device > which is already running. > > My configuration uses manual and not dhcp: > > iface eth0 inet manual > > 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 reset to manual as a precaution; I didn't think it mattered much since if you touch the interface in a diskless setting you're hosed. But your testing shows manual is safer. > > This nfs startup part is a part that seems to have suffered from the > transition from boot time scripts to event driven scripts. This kind > of thing use to work in the previous init script way. I don't know > the best design to make this work in the general case for the new > event driven way. But at least in my testing it was sufficient to add > this mount line to /etc/rc.local in order to mount additional nfs > mount points at boot time. Do you think this is something that merits a bug/wish report? Against what package? > > File /etc/rc.local: > > mount -a -t nfs Thanks; I can never remember where the "do this after startup" stuff goes. I did this and it worked. I tried Michael's suggestion to use _netdev, before changing rc.local, but it didn't help. The man says its for things that have to happen after network access is available. On the one hand, network access is always available in a diskless setup; on the other hand, the regular startup sequence never gets an indication that the network has come up, which I guess means neither _netdev not nfs mounts fire. > > In my setup when I added that then everything seemed to work just > fine. Because nfsdiskless isn't one of the really standard > configurations, you need to set it up manually, I think that is > probably a good enough solution for the moment. Give it a go. Definitely. It was really annoying have to do the mount manually every time, and if my daughter starts the machine she can't do it, and so can't use myth. Why I need the nfs mount for myth I don't know; the system is a client, getting recordings from the server (the same machine serving NFS). So it seems to me it should only matter if the myth server can see the files. But it doesn's work that way. Thanks also to Roger for his suggestions about tempfs; I'm going to look into them. I got them from something I copied from the net, and so they may well be inappropriate for the latest Debian. Ross > > Bob -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/1355087373.5409.13.ca...@corn.betterworld.us
