At 22:02 19/07/2005, you wrote:
No, kernel NFS can be either compiled or loaded as a module into a custom
kernel.
Does anyone know which NFS server I *should* be using? Does anyone know why
lockd fails to run?
But the userspace daemons
(/sbin/rpc.lockd and /sbin/rpc.statd) don't seem to want to start even
when I call them from the command line. A quick grep through my kernel
config suggests this is because lockd is supposedly built into the kernel;
tamora:/usr/src/linux# cat .config | grep -i lockd
CONFIG_LOCKD=y
CONFIG_LOCKD_V4=y
My kernel just shows the first option, and I'm not using the kernel server.
Although I can't for the life of me find the option in menuconfig (maybe
due to me copying the .config over from my old 2.4 installation...? But I
can't find the option in 2.4's menuconfig either).
I found it by enabling the "Show All Options" option while running "make
xconfig."
Using that option I also found that in order to enable CONFIG_LOCKD_V4, I
first
had to enable NFS server support (NFSD) and then NFSv3.
Well, I don't have X on this machine but unchecking the NFS modules does in
turn uncheck the lockd module too. If I'm using the userspace server,
should I remove the in-kernel NFS server components?
Aha, just had another look through menuconfig and found the root of the
problem; having NFSD enabled in the kernel stops the userspace server from
loading lockd (that'll teach me not to read the help text properly). Just
switched back to another kernel and using the kernel server, and both lockd
and statd both seem to be running OK, and no more locking errors on the
clients.
Unfortunately, it still hasn't fixed my other ongoing problem
http://lists.debian.org/debian-user/2005/06/msg03111.html with symlinks not
working on exported filesystems, but it's a step in the right direction...!
Thanks for your help Marty, it set me on the right track!
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]