Bug#746358: Boot hangs if fstab contains NFS entries

[Michael Biebl]

Hi Daniel,

Am 04.05.2014 11:20, schrieb Daniel Leidert:
> I also have some NFS entries in my /etc/fstab and I manually
> patched /etc/network/if-up.d/mountnfs as suggested after I upgraded the
> system. However: the system still hangs on startup for 1-2 minutes
> saying:
> 
> [ OK ] Started LSB: NFS support files common to client and server
> [  **] (X of Y) A start job is running for <location>
> 
> The good thing though is, it is not a deadlock. But of course I don't
> want to wait several minutes for my system to boot. Your patch doesn't
> fix *this* issue.
> 
> I'm using the network-manager package to connect to my local network.

The issue you are seeing is a slightly different one. I'll try to
explain the reasons for it.

We switched to native remote mount handling (previously, we relied on
the mountnfs ifup.d hook for that).

We also added support for LSB defined facilities in 204-9.
That means, with 204-9, the /etc/init.d/networking init script will
activate the network.target.

As soon as that target is active, systemd will mount the remote file
systems defined in /etc/fstab.

In your case, you do not actually use ifupdown to bring up the network.
As a result, when systemd tries to mount the NFS share, the network is
not yet up and the mount request will time out.

Unfortunately, we can't easily make NetworkManager.service pull in
network.target, because with our current support for rcS type services
this would create a dependency loop.

I would suggest, that if you are using NetworkManager, you switch your
NFS mount to use comment=systemd.automount to delay the actual mount
request until it is accessed.

Regards,
Michael




-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 901 bytes
Desc: OpenPGP digital signature
URL: <http://alioth-lists.debian.net/pipermail/pkg-systemd-maintainers/attachments/20140504/064ab49b/attachment-0002.sig>