Received-SPF: pass (google.com: domain of linux-nfs+bounces-696-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
From: Benjamin Coddington <bcodding@redhat.com>
To: Martin Wege <martin.l.wege@gmail.com>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: umount.nfs: /mnt: block devices not permitted on fs
Date: Mon, 18 Dec 2023 16:43:17 -0500
Message-ID: <F17FB28B-0A94-4595-AEC6-0C1A53BAC8B2@redhat.com>
In-Reply-To: <CANH4o6NcrFz6EKDcfq1daaSpmOuCebYA-A3u3Bk_+kPtJWBLYA@mail.gmail.com>
References: <CANH4o6OdF=hjLtZ1_jbqPjexuenYn6cSvxFrJ+BUUDQv86DRzA@mail.gmail.com>
 <CANH4o6Mu8kDvZTrssVc_Tr1CKkWaUFnZxzdjQNw7-2tmYTTOzw@mail.gmail.com>
 <7849E95F-603B-49B8-B251-73CAB4811E48@redhat.com>
 <CANH4o6NcrFz6EKDcfq1daaSpmOuCebYA-A3u3Bk_+kPtJWBLYA@mail.gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

On 18 Dec 2023, at 3:37, Martin Wege wrote:

> On Thu, Nov 30, 2023 at 1:26 PM Benjamin Coddington <bcodding@redhat.com> wrote:
>>
>> On 29 Nov 2023, at 22:45, Martin Wege wrote:
>>
>>> ?
>>>
>>> On Fri, Nov 24, 2023 at 9:57 AM Martin Wege <martin.l.wege@gmail.com> wrote:
>>>>
>>>> Hello,
>>>>
>>>> We get a umount.nfs: /mnt: block devices not permitted on fs in a snap
>>>> container on Ubuntu.
>>>>
>>>> Can anyone explain what is going wrong?
>>
>> Something in umount is calling umount_error() with -EACCES.  I admit that
>> error message makes no sense, it seems to be cruft.
>
> Can this be fixed, please?

Probably, have to figure out the problem first.

>> Maybe you can run the umount in gdb with a breakpoint on umount_error() and
>> send along the stack.
>
> This is much more difficult than I expected - the target system is a
> Ubuntu Core system, which uses SNAPs, and debugging there is a
> disaster.

Hm, I don't know what SNAPs are..

> Does umount have an extended debug output?

Not really, maybe try with "strace" and we can see what the kernel's
returning?  From there, it might look like flipping on some debugging (check
out the dprintk's in fs/nfs and net/sunrpc and the `rpcdebug` userspace
tool), or maybe some tracepoints would help.

Ben