Fedora 20 nfs

[T.C. Hollingsworth]

On Tue, Oct 15, 2013 at 8:33 AM, Mike Chambers <mike at mtchambers.com> wrote:
> Tried your script but got an exit code, ugh can't remember it now.

That was my fault, sorry.  The quoting was screwed up and an escape
character was transposed.  (I blame Monday.  :-p)

The correct line should be:

ExecStart=/bin/bash -c "until ifconfig p5p1 | grep '192\.168'; do
continue; done"

(Also make sure it's all on one line despite my mailer's insistence
that it should be two.  ;-)

Also, come to think of it, skipping right to a ping test might be most accurate:

ExecStart=/bin/bash -c "until ping -c1 8.8.8.8; do continue; done"

Feel free to swap out Google's IP address for your NFS server or
something if you'd like.

>> 3.) NM-wait-online is working fine, but systemd isn't waiting for it
>> for some reason.
>
> I think above is the problem, but can't swear to it.'
sti
Actually, looking at the log you posted in that bug, NM-wait-online is
in fact returning successfully before systemd attempts to mount your
NFS share, so it is indeed one of:

1.) The `nm-online` tool lies.

2.) NFS is failing for some other random reason other than the network
not being up.

Either way, it's not a systemd bug.  Fixing really-wait-online.service
as described above will narrow down whether it's 1 or 2.  If it's 1,
you'll already have a temporary fix working and can complain to the
NetworkManager folks.  If NFS mounts still fail, it is most definitely
not the network, so file a bug against nfs-utils and let them try and
troubleshoot their obtuse error output.

> Got confused on your above temp fix.  So adding home-download.mount to the end of the Before line, in my wait.online service file?

Shouldn't be necessary.  As I said above, your logs indicate systemd
is in fact doing what it is supposed to without this hack.

-T.C.