Bug#622394: systemd: nfs-common and rpcbind unit files to fix systemd NFS issues properly

[Simon McVittie]

Control: clone 622394 -2 -3
Control: retitle -2 NFS mounts from /etc/fstab do not work
Control: submitter -2 matt at mattgrant.net.nz
Control: retitle -3 NFS exports fail due to rpcbind not starting before nfs-common and nfs-kernel-server
Control: submitter -3 matt at mattgrant.net.nz
Control: tags -3 moreinfo unreproducible
Control: user release.debian.org at packages.debian.org
Control: usertag -2 - jessie-is-blocker
Control: usertag -3 - jessie-is-blocker
Control: tags 622394 unreproducible

On 16/01/15 21:27, Niels Thykier wrote:
> Can anyone provide me with an up to date status on this bug?  It seems
> to be a mismatch of no less than 3 different bugs if I understand
> Simon's comment in message #41 correctly (see [1] for the full text).

That is my understanding, but I am not an expert on this package (I
don't use NFS myself).

I was hoping for a response from a maintainer and/or bug reporters,
but apparently I'm not going to get one, so... cloning it is. I hope
I got the usertags and stuff right - please check.

> In particular, this bug is correctly marked as a Jessie blocker due to
> the current bug title, which suggests that nfs-common has a circular
> dependency causing NFS to break under systemd.  However, reading the bug
> I am understanding that this loop is now assumed to be fixed?

That's my "bug A" and yes I suspect it's been fixed by better
sysvinit compatibility in systemd... but I could be wrong, and I
don't know specifically what the fix was. Let's use #622394 for this.

I haven't closed it immediately, but I've tagged it unreproducible.
I certainly didn't see it myself.

>  If the other bugs (see below) are still present, let us clone this
> bug/file separate bugs for those!

I believe M1 and M2 may be separate bugs, and C1 might be the same
thing as M1 or a fourth bug (it's unclear).

The clone "NFS mounts from /etc/fstab do not work" is M1 and probably
also C1; I'm leaving them entangled for now. Matthew, Chris, if
you have any additional info / newer test results / etc. on this
specific issue, please send it to that clone (check
https://bugs.debian.org/nfs-common for its bug number).

The clone "NFS exports fail due to rpcbind not starting before
nfs-common and nfs-kernel-server" is M2. Matthew, if you have
additional info on that, please send it there (again, check
https://bugs.debian.org/nfs-common for its bug number).

> On the topic of "_netdev": Did you conclude that systemd correctly
> auto-discovered NFS as a _netdev *without* explicitly stating it in
> /etc/fstab?

I believe so. systemd does have code to do this, and Ben's logic
concluding that the test results indicate it is working
correctly does seem valid.

    S