Rich Freeman on 22 Jan 2014 19:55:53 -0800


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] network service errors


On Wed, Jan 22, 2014 at 4:26 PM, JP Vossen <jp@jpsdomain.org> wrote:
> On 01/22/2014 03:57 PM, Carl Johnson wrote:
>>
>> still can't get the nfs volume to mount on boot. fstab looks like this...
>>
>> fileserver:/drive_z_bak  /backups                nfs
>> _netdev,defaults,auto,noatime,intr              0 0
>>
>> "mount -a" makes it mount right up. just not on boot. why?
>
>
> I've had lots of timing issues with NFS (on Debian).  I finally gave up and
> brute-forced it:

Must be some kind of curse.  I was struggling with getting an NFS
share to work correctly with systemd on Gentoo.  Despite tagging the
fstab entry with _netdev or whatever it would still try to mount it
before running dhchcd, which obviously doesn't work.  I think I too
ended up hacking in pre/post sleep commands in a few places.
Automount worked a little better, but I had a service that needed the
share and automount wasn't quite fast enough for it.

I chalked that up to systemd being immature.  I think the bigger issue
is that NFS doesn't seem to be all that common and well-tested in
general.  Linux 3.12 breaks NFS on my server for a reason I have yet
to decipher (but only if I have eth0 bridged for use with KVM - and I
can't get the same behavior inside a VM for testing).  Linux 3.10
breaks NFS under the same circumstances unless I enable NFS v4 on
server and clients.  Oh, and my NFS root Gentoo box won't shut down
correctly after an update a year or so ago (not that big of a deal -
not going to lose data on a diskless system).

Oh, and if you have an old ATI HD PCI tuner card 3.12 might break that
too...  I have a patch from the maintainer if you need it (I think it
is going to get merged one of these releases)...

Looks like I'll get to spend a weekend with wireshark one of these
days - can't stay on 3.10 forever.  The obvious debugging tricks like
running rpcinfo from a client don't turn up much.  Besides, listeners
typically bind to IPs, and the IP of my bridge is the same static
address eth0 used to have, so what gives?

Rich
___________________________________________________________________________
Philadelphia Linux Users Group         --        http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion  --   http://lists.phillylinux.org/mailman/listinfo/plug