[Buildroot] [PATCH 1/1 v1] nfs-utils: install /etc/nfsmount.conf

[Jonathan Ben Avraham]

On Tue, 22 Dec 2015, Yann E. MORIN wrote:

> Date: Tue, 22 Dec 2015 19:51:31 +0100
> From: Yann E. MORIN <yann.morin.1998 at free.fr>
> To: Jonathan Ben Avraham <yba at tkos.co.il>
> Cc: buildroot at busybox.net
> Subject: Re: [Buildroot] [PATCH 1/1 v1] nfs-utils: install /etc/nfsmount.conf
> 
> Jonathan, All,
>
> On 2015-12-22 19:48 +0200, Jonathan Ben Avraham spake thusly:
>> On Tue, 22 Dec 2015, Yann E. MORIN wrote:
>>> On 2015-12-22 19:12 +0200, Jonathan Ben-Avraham spake thusly:
>>>> From: Jonathan Ben Avraham <yba at tkos.co.il>
> [--SNIP--]
>>>> +	$(INSTALL) -D -m 644 \
>>>> +		$(@D)/utils/mount/nfsmount.conf $(TARGET_DIR)/etc/nfsmount.conf
>>>
>>> This file is only comments, save for a single un-commented line;
>>>   [ NFSMount_Global_Options ]
>>>
>>> Would it not be better that we just generate a file with just that line,
>>> and a comment to see the sources for the format, like so:
>>>
>>>   # See utils/mount/nfsmount.conf in nfs-utils sources
>>>   # for the format of this file.
>>>   [ NFSMount_Global_Options ]
>>
>> When debugging an unexpected NFS mount problem on a target for which you do
>> not have the build system, at a site that might not have Internet access,
>> with a root filesystem that might not have the "strings" utility, it is much
>> easier to have all of the mount options in front of you rather than to have
>> to leave the premises and start searching for the correct nfs-utils sources
>> in order to find the full nfsmount.conf. It's just a much more friendly way
>> of working. The lack of any nfsmount.conf cost me (my client really) several
>> hours of work while I installed "strings" and figured out that there is such
>> a thing as nfsmount.conf. It would be unfortunate for others to have to
>> repeat this. Most people assigned to debug these types of systems are even
>> less experienced than I.
>
> Well, one would argue that the system should have been tested before
> being deployed. That's DevOps in action! Hehe! ;-)  (Kidding!)

Hi Yann,
Yes, they usually are tested before deployment, but often the end-user 
environment changes in ways that the testers did not expect and the 
default values of unusual parameters need to be changed. I think that NFS 
mount is particularly problematic in this regard because of the many 
different ways that sys-admins set it up and the wide range of servers 
used.

> But as I said: I'm OK with the full file as well. If a user is so
> concerned with the size, he can provide a post-build script to tweak
> that file, or a rootfs-overlay to provide his own version.
>
> So, my Reviewed-by stands. ;-)

tnx,

  - yba


>
> Thanks!
>
> Regards,
> Yann E. MORIN.
>
>

-- 
  9590 8E58 D30D 1660 C349  673D B205 4FC4 B8F5 B7F9  ~. .~  Tk Open Systems
=}-------- Jonathan Ben-Avraham ("yba") ----------ooO--U--Ooo------------{=
mailto:yba at tkos.co.il tel:+972.52.486.3386 http://tkos.co.il skype:benavrhm