Bug#804241: Please update initramfs in postinst (maybe)

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

Bug#804241: Please update initramfs in postinst (maybe)

Steve McIntyre
Source: f2fs-tools
Version: 1.4.1-1
Severity: important
Tags: d-i

Hi!

Since the move to systemd as the default init system, the initramfs
will attempt to fsck and mount both / and /usr (where applicable). To
aid this, initramfs-tools will copy necessary filesystem tools into
the initramfs when it is generated.

To make this work well, all filesystem tools packages for filesystems
that are likely to be used for / and/or /usr should call
"update-initramfs -u" in their postinst. This will

 (a) ensure that necesssary fsck tools are included in the initramfs
     generated by debian-installer (see #801961 for an example failure
     here); and
 (b) ensure that bug fixes to fsck tools get included immediately in
     the initramfs

I've checked your package and I don't see any update-initramfs
calls. Please add one, if you consider f2fs to be a likely/sensible
option as a base (/ or /usr) filesystem. If you'd like help doing that
postinst work, I can supply a patch - just ask!

-- System Information:
Debian Release: 8.2
  APT prefers stable
  APT policy: (500, 'stable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 3.16.0-4-amd64 (SMP w/4 CPU cores)
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Reply | Threaded
Open this post in threaded view
|

Bug#804241: Please update initramfs in postinst (maybe)

Vincent Cheng-2
Hi Steve,

On Fri, Nov 6, 2015 at 5:54 AM, Steve McIntyre <[hidden email]> wrote:

> Source: f2fs-tools
> Version: 1.4.1-1
> Severity: important
> Tags: d-i
>
> Hi!
>
> Since the move to systemd as the default init system, the initramfs
> will attempt to fsck and mount both / and /usr (where applicable). To
> aid this, initramfs-tools will copy necessary filesystem tools into
> the initramfs when it is generated.
>
> To make this work well, all filesystem tools packages for filesystems
> that are likely to be used for / and/or /usr should call
> "update-initramfs -u" in their postinst. This will
>
>  (a) ensure that necesssary fsck tools are included in the initramfs
>      generated by debian-installer (see #801961 for an example failure
>      here); and
>  (b) ensure that bug fixes to fsck tools get included immediately in
>      the initramfs
>
> I've checked your package and I don't see any update-initramfs
> calls. Please add one, if you consider f2fs to be a likely/sensible
> option as a base (/ or /usr) filesystem. If you'd like help doing that
> postinst work, I can supply a patch - just ask!

Yep, this sounds like a valid use case (e.g. my Raspberry Pi uses a
f2fs-formatted root filesystem).

All that's needed here is an unconditional call to "update-initramfs
-u" on postinst and postrm, right? Just to make sure I get this right,
would you be able to point me to another package that does the right
thing here (or even a patch if you have time)? Thanks!

Regards,
Vincent

Reply | Threaded
Open this post in threaded view
|

Bug#804241: Please update initramfs in postinst (maybe)

Steve McIntyre
On Thu, Nov 19, 2015 at 03:23:59AM -0800, Vincent Cheng wrote:

>>
>> I've checked your package and I don't see any update-initramfs
>> calls. Please add one, if you consider f2fs to be a likely/sensible
>> option as a base (/ or /usr) filesystem. If you'd like help doing that
>> postinst work, I can supply a patch - just ask!
>
>Yep, this sounds like a valid use case (e.g. my Raspberry Pi uses a
>f2fs-formatted root filesystem).
>
>All that's needed here is an unconditional call to "update-initramfs
>-u" on postinst and postrm, right? Just to make sure I get this right,
>would you be able to point me to another package that does the right
>thing here (or even a patch if you have time)? Thanks!

Sure, no problem. As you've guessed, what I've done for other packages
is trivial - see the debdiff in #804247 for reiserfsprogs as an
example.

--
Steve McIntyre, Cambridge, UK.                                [hidden email]
"I can't ever sleep on planes ... call it irrational if you like, but I'm
 afraid I'll miss my stop" -- Vivek Das Mohapatra