update-initramfs -k all -u
Hi, I maintain uswsusp. It is a package that relies on a binary on the initramfs that will start the resume process. This binary (and some other stuff) get installed via an update-initramfs call in the postinst. On some updates, the new binary that suspends the system is incompatible with the old resume binary on the initramfs, hence a update-initramfs call is required. So far so good. Current pratice is to only call `update-initramfs -u', that is, to only update the most recent initramfs. This however will break older initramfses (I have had bugreports). Now I plan to switch to running with '-k all' (all initramfses), but this of course will also influence the packages that ran with '-u' only. Now what do people think is the best option? (And why?) grts Tim signature.asc Description: PGP signature
Re: update-initramfs -k all -u
> Hi, > > I maintain uswsusp. It is a package that relies on a binary on the > initramfs that will start the resume process. This binary (and some > other stuff) get installed via an update-initramfs call in the postinst. > On some updates, the new binary that suspends the system is > incompatible with the old resume binary on the initramfs, hence a > update-initramfs call is required. > > So far so good. > > Current pratice is to only call `update-initramfs -u', that is, to only > update the most recent initramfs. This however will break older > initramfses (I have had bugreports). Now I plan to switch to running > with '-k all' (all initramfses), but this of course will also influence > the packages that ran with '-u' only. > > Now what do people think is the best option? (And why?) This may be related: http://lists.debian.org/debian-devel/2006/12/msg00014.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: update-initramfs -k all -u
also sprach Tim Dijkstra <[EMAIL PROTECTED]> [2007.05.15.2201 +0200]: > Now what do people think is the best option? (And why?) I use -k all in mdadm already. I could not find any reasons why that would not be a good idea. -- Please do not send copies of list mail to me; I read the list! .''`. martin f. krafft <[EMAIL PROTECTED]> : :' : proud Debian developer, author, administrator, and user `. `'` http://people.debian.org/~madduck - http://debiansystem.info `- Debian - when you have better things to do than fixing systems when everything is coming your way, you're in the wrong lane. signature.asc Description: Digital signature (GPG/PGP)
Re: update-initramfs -k all -u
On Wed, May 16, 2007 at 12:23:55AM +0200, martin f krafft wrote: also sprach Tim Dijkstra <[EMAIL PROTECTED]> [2007.05.15.2201 +0200]: Now what do people think is the best option? (And why?) I use -k all in mdadm already. I could not find any reasons why that would not be a good idea. The reason I didn't use it in cryptsetup's initramfs script was that many people who have > 1 kernel use the second one as a fallback. In case the updated version of the package which calls update-initramfs in its postinst contains some grave bug which renders the initramfs image useless, the old one will still be there... -- David Härdeman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: update-initramfs -k all -u
also sprach David Härdeman <[EMAIL PROTECTED]> [2007.05.18.0118 +0200]: > The reason I didn't use it in cryptsetup's initramfs script was that > many people who have > 1 kernel use the second one as a fallback. In > case the updated version of the package which calls update-initramfs in > its postinst contains some grave bug which renders the initramfs image > useless, the old one will still be there... update-initramfs does leave a backup file around... -- Please do not send copies of list mail to me; I read the list! .''`. martin f. krafft <[EMAIL PROTECTED]> : :' : proud Debian developer, author, administrator, and user `. `'` http://people.debian.org/~madduck - http://debiansystem.info `- Debian - when you have better things to do than fixing systems NP: Man on Fire | Habitat signature.asc Description: Digital signature (GPG/PGP)
Re: update-initramfs -k all -u
Am Freitag 18 Mai 2007 08:26 schrieb martin f krafft: > also sprach David Härdeman <[EMAIL PROTECTED]> [2007.05.18.0118 +0200]: > > The reason I didn't use it in cryptsetup's initramfs script was that > > many people who have > 1 kernel use the second one as a fallback. In > > case the updated version of the package which calls update-initramfs in > > its postinst contains some grave bug which renders the initramfs image > > useless, the old one will still be there... > > update-initramfs does leave a backup file around... Does that help you when you cannot boot anymore? HS pgp4CeKacfdhG.pgp Description: PGP signature
Re: update-initramfs -k all -u
On Fri, May 18, 2007 08:26, martin f krafft wrote: > also sprach David Härdeman <[EMAIL PROTECTED]> [2007.05.18.0118 +0200]: >> The reason I didn't use it in cryptsetup's initramfs script was that >> many people who have > 1 kernel use the second one as a fallback. In >> case the updated version of the package which calls update-initramfs in >> its postinst contains some grave bug which renders the initramfs image >> useless, the old one will still be there... > > update-initramfs does leave a backup file around... True, but that won't help if two initramfs-tools using packages have been updated since last boot and it also requires the user to know that the backup exists and how to tell grub to use it. -- David Härdeman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: update-initramfs -k all -u
also sprach Hendrik Sattler <[EMAIL PROTECTED]> [2007.05.18.0842 +0200]: > Does that help you when you cannot boot anymore? grub provides an interactive shell. also sprach David Härdeman <[EMAIL PROTECTED]> [2007.05.18.0844 +0200]: > True, but that won't help if two initramfs-tools using packages have been > updated since last boot and it also requires the user to know that the > backup exists and how to tell grub to use it. Good point. Thus it looks like there is no right way other than not to introduce bugs into initramfs-producing packages. Not using -k all may mean that your existing initramfs cannot boot the current system anymore, and using it may mean that your previous and only good initramfs now also gets infected by the I-won't-let-you-boot-no-more virus. -- Please do not send copies of list mail to me; I read the list! .''`. martin f. krafft <[EMAIL PROTECTED]> : :' : proud Debian developer, author, administrator, and user `. `'` http://people.debian.org/~madduck - http://debiansystem.info `- Debian - when you have better things to do than fixing systems "one should never trust a woman who tells her real age. if she tells that, she will tell anything." -- oscar wilde signature.asc Description: Digital signature (GPG/PGP)
Re: update-initramfs -k all -u
> "David" == David Härdeman <[EMAIL PROTECTED]> writes: David> True, but that won't help if two initramfs-tools using David> packages have been updated since last boot and it also David> requires the user to know that the backup exists and how to David> tell grub to use it. ...or if using a platform not supported by Grub. Including all PowerPC systems for example. -- Brian May <[EMAIL PROTECTED]>
Re: update-initramfs -k all -u
On Fri, 18 May 2007 01:18:01 +0200 David Härdeman <[EMAIL PROTECTED]> wrote: > On Wed, May 16, 2007 at 12:23:55AM +0200, martin f krafft wrote: > >also sprach Tim Dijkstra <[EMAIL PROTECTED]> [2007.05.15.2201 +0200]: > >> Now what do people think is the best option? (And why?) > > > >I use -k all in mdadm already. I could not find any reasons why that > >would not be a good idea. > > The reason I didn't use it in cryptsetup's initramfs script was that > many people who have > 1 kernel use the second one as a fallback. In > case the updated version of the package which calls update-initramfs in > its postinst contains some grave bug which renders the initramfs image > useless, the old one will still be there... Yes, but when initramfs and the root partition need to be in sync to be functional, the backup won't help you either... grts Tim signature.asc Description: PGP signature
Re: update-initramfs -k all -u
On Fri, May 18, 2007 at 09:12:15AM +0200, martin f krafft wrote: > Good point. Thus it looks like there is no right way [...] How about making the use of "-k all" configurable? Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences -
Re: update-initramfs -k all -u
On Fri, 18 May 2007 16:41:41 +0200 Gabor Gombas <[EMAIL PROTECTED]> wrote: > On Fri, May 18, 2007 at 09:12:15AM +0200, martin f krafft wrote: > > > Good point. Thus it looks like there is no right way [...] > > How about making the use of "-k all" configurable? That seems like a good idea... but what would be the default? grts Tim signature.asc Description: PGP signature
Re: update-initramfs -k all -u
On Fri, May 18, 2007 at 04:51:46PM +0200, Tim Dijkstra wrote: > That seems like a good idea... but what would be the default? Doesn't matter as there are already examples of both behavior so picking a random value and flamdebating it later is fine. The important thing is to make the behavior consistent between packages and allow people who are unhappy with the default to override it. Gabor -- - MTA SZTAKI Computer and Automation Research Institute Hungarian Academy of Sciences -
Re: update-initramfs -k all -u
On Fri, 18 May 2007 19:25:09 +0200 Gabor Gombas <[EMAIL PROTECTED]> wrote: > On Fri, May 18, 2007 at 04:51:46PM +0200, Tim Dijkstra wrote: > > > That seems like a good idea... but what would be the default? > > Doesn't matter as there are already examples of both behavior so picking > a random value and flamdebating it later is fine. The important > thing is to make the behavior consistent between packages and allow > people who are unhappy with the default to override it. OK, Just implemented the debconf question, see 425050. If it gets applied I'll file some patches for packages using initramfs-tools to use the same question. grts Tim signature.asc Description: PGP signature
Re: update-initramfs -k all -u
On Fri, May 18, 2007 at 04:41:41PM +0200, Gabor Gombas wrote: On Fri, May 18, 2007 at 09:12:15AM +0200, martin f krafft wrote: Good point. Thus it looks like there is no right way [...] How about making the use of "-k all" configurable? The point I was trying to make before is that the decision whether "-k all" is sane or not can only be decided per package (and perhaps even per revision). For example, a bad cryptsetup binary written to all available initramfs images means no boot at all. On the other hand, if a security bug would be found in the cryptsetup initramfs scripts, "-k all" would be quite necessary. -- David Härdeman -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: update-initramfs -k all -u
On Sat, 19 May 2007 00:23:08 +0200 David Härdeman <[EMAIL PROTECTED]> wrote: > On Fri, May 18, 2007 at 04:41:41PM +0200, Gabor Gombas wrote: > >On Fri, May 18, 2007 at 09:12:15AM +0200, martin f krafft wrote: > > > >> Good point. Thus it looks like there is no right way [...] > > > >How about making the use of "-k all" configurable? > > The point I was trying to make before is that the decision whether "-k > all" is sane or not can only be decided per package (and perhaps even > per revision). That sounds reasonable, but unfortunately you can't. If I (as the uswsusp maitainer) decide to use '-k all' your package gets installed to the initramfs anyway. So this is really something we should agree on for all packages at same time. As it seems we can't agree that's why I think using the users preference is the best we can do. BTW, I was reading about `triggers' on the dpkg-list a while back. Update-initramfs seems like a perfect use case. Updating all initramfses three times in one dpkg run, just because initrfamfs-tools, uswsusp and cryptsetup get updated at the same time is not funny... grts Tim signature.asc Description: PGP signature
Re: update-initramfs -k all -u
Tim Dijkstra wrote: >Content-Type: text/plain; charset=UTF-8 >Content-Transfer-Encoding: quoted-printable You should fix your mail system not to send quoted-printable. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: update-initramfs -k all -u
On Sat, May 19, 2007 at 06:31:02PM +0300, Touko Korpela wrote: >> Content-Type: text/plain; charset=UTF-8 >> Content-Transfer-Encoding: quoted-printable > You should fix your mail system not to send quoted-printable. Urm, why? /* Steinar */ -- Homepage: http://www.sesse.net/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]