On 17/02/17 21:13, Kevin Lemonnier wrote:
At least on Linux the man page for mount() seems to indicate that it'll
be enough, as long as the kernel as support for it.
If 1 and 2 are already supported by the kernel itself, that should just
work I think ?
Normally sort of, but NFS is not normal. NFS requires interacting with
rpcbind to figure out which port mountd is running on (*), and then with
mountd to get the file system handle for the root node. That file
handle is then passed to the driver. So, there's quite a bit you need
to do before you call mount() for NFS.
*) IIRC "webnfs" runs on a constant port and doesn't require the rpc
tomfoolery
I could use the mount_nfs through sysproxy to call mount() on the instance,
is that correct ? Would be good to test it out without having to modify
the actual code in config.c and the rumprun script to support NFS yet.
yes
At least glusterfs might work using the same principle, except there are
a lot more details to consider and places to hang yourself.
When you say it like that .. :)
I say it like that because that's how it is.
