Bug#724275: initramfs-tools should change the resume device

Rodolfo García Peñas kix at debian.org
Fri Nov 1 12:30:53 UTC 2013


Hi Olivier,

1. First, there is a problem with the typo in the initramfs-tools package. This problem will be solved in next version of initramfs-tools, because the maintainers discovered this problem previosly (see http://anonscm.debian.org/gitweb/?p=kernel/initramfs-tools.git;a=blobdiff;f=hooks/resume;h=96e2ffdedea693ff3f5d1e287eef3ad58850d20a;hp=8554fb2d5885b7b1cd1f4aa8de97717340dd6fd1;hb=736200b04b0974b14b59719220a0ba77da57f28d;hpb=502b1f84aaaf76f0d96502532e991dc71673235f). Ben/Maximilian, do you have in mind upload a new version of initramfs-tools soon?

So, if we change the typo (initramfs-tools maintainers upload a new version), and the user don't use LVM2, the problem is gone.

2. Second, if we solve the problem with the typo, and we use LVM2 (like your system), the system won't boot. In this case is because the LVM2 scripts cannot check if the filesystem in format UUID= is the swap in a LVM volume. So, the problem is in the initramfs-tools lvm2 script, because is unable to know it. This problem was created by the latest version of uswsusp package, because it use now UUID devices (no /dev devices), but the problem should be solved in lvm2, because is unable to handle UUID devices. This problem is the bug number #678687 in lvm2. IMO, this bug should be forwareded/merged with #678687 when the problem with the typo will be solved. So, I don't have problems to hold the bug in uswsusp package if the initramfs-tools package is updated soon, and then forward it to lvm2.

So, the next step is try to make a patch for lvm2 package to handle UUID devices.

This problem is critical, because the system don't boot if uswsusp is installed, but uswsusp is ok.

Olivier, do you agree?

Best regards,
kix

On Wed, 30 Oct 2013, Olivier Berger escribió:

> On Wed, Oct 30, 2013 at 03:25:51PM +0100, Olivier Berger wrote:
> > 
> > I've seen the suggestion made by Kapil Hari Paranjape in
> > http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=725275#27 which seems
> > to give some guidance, but it's not really clear for me so far.
> > 
> 
> I've investigated this a bit, and found out that by setting "resume device = /dev/mapper/main-swap" in /etc/uswsusp.conf, the problem now seems solved.
> 
> # grep -i resume /etc/uswsusp.conf /etc/initramfs-tools/conf.d/resume
> /etc/uswsusp.conf:resume device = /dev/mapper/main-swap
> /etc/initramfs-tools/conf.d/resume:RESUME=/dev/mapper/main-swap
> 
> So, provided that patch from #725275 is applied to fix the blkid typo, and there's a consistent /dev/mapper/VG-swap like setting in /etc/uswsusp.conf, everything seems to work fine.
> 
> Btw, Kix, did you try to reproduce with initramfs-tools v. 0.114, cause in 0.113, it seemed to be working, according to #725275.
> 
> Now, I tend to think that we have a solution, in that uswsusp shouldn't use /dev/dm-N like paths.
> 
> Hope this helps.
> 
> Best regards,
> -- 
> Olivier BERGER 
> http://www-public.telecom-sudparis.eu/~berger_o/ - OpenPGP-Id: 2048R/5819D7E8
> Ingenieur Recherche - Dept INF
> Institut Mines-Telecom, Telecom SudParis, Evry (France)
> 

-- 
 .''`.  Rodolfo García Peñas (kix) <kix at debian.org>
: :'  : Proud Debian Developer
`. `'`  4096R / 3F48 0B8C C385 AD41 9E28  006A 7B1F 5490 72B7 4923
 `-     Debian - when you have better things to do than fixing systems
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 836 bytes
Desc: Digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-lvm-maintainers/attachments/20131101/231f5509/attachment.sig>


More information about the pkg-lvm-maintainers mailing list