Bug#693485: keyboard-configuration: XKBLAYOUT, XKBOPTIONS values manually set in /etc/default/keyboard lost on upgrade
On Sat, Nov 17, 2012 at 04:15:06AM +0400, Bob Bib wrote: > > after upgrading the whole 'console-setup' package from 1.82 to 1.87 > version, I've lost my XKBLAYOUT and XKBOPTIONS variable values > manually set in '/etc/default/keyboard'; Can you give us the following information: 1. The original values of XKBLAYOUT and XKBOPTIONS. 2. The changed values of XKBLAYOUT and XKBOPTIONS. 3. Whether this bug is reproducible or console-setup no longer overwrites your manual settings. Anton Zinoviev -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121120113355.ga4...@logic.fmi.uni-sofia.bg
Re: Proposal to add patches to netcfg (#682737)
Hi, so your patches looked very good, and as Michael already mailed, they have been merged (but not yet uploaded). I need to try yet another installation with the desktop task activated, which takes a fair while. Afterwards it should be uploaded, hit the dailies and hopefully not be the cause of much breakage. ;-) I did tweak a couple of things: * e946cee nm-conf: Use Linux's random/uuid proc entry if available. * ba5ee42 Write out network-manager configuration files with mode 0600. * 46208b2 Reword target_network_config template; do not ask the question. * d2dd5a1 nm-conf: use ESSID as connection name * 83e9215 test: reactivate test_nc_v6_interface_configured_suite * 4e1dbbd finish-install: Do not set /etc/network/interfaces to mode 0600. * 5bfce24 nm-conf: Allow both IPv4 and IPv6 to be activated post-installation. * 7f9f1d2 nm-conf: Only write out MAC addresses on static configurations. * c1e4697 Add a changelog entry for the n-m write config changes. So we need libuuid only on non-Linux platforms (I hope it still compiled there). As discussed the mode of both n-m config files and /e/n/i was adjusted to be private to root. The ESSID connection name change is done because Michael said that n-m's default changed, hence do the same in netcfg. I longly discussed the MAC write out with Michael and I guess we reached a sane conclusion here that still allows e.g. images be produced with one hardware (say KVM) and reused on another one while keeping the settings and not creating a new connection while still dealing with multiple NICs. The IPv4/IPv6 change is IPv6 enablement: if IPv6 gets activated on the network at some point later post-installation the machine should still take advantage of it instead of ignoring it forever. Two tiny notes about the coding style: * You aligned struct members horizontally. That causes diffs to be more noisy, makes them more likely to conflict and more lines need to be changed when a new member is added. * You passed the structs in the nm_write_* function by value. Is there a reason for this? I presume it's partly due to the fact that nm_config_info inlines the structs. Still the structure's parts should not be copied when passed to another function, hence pointers should be used. I did not refactor this when basing my changes upon your branch, though. Lastly thanks for following up on this and your great work. :-) Kind regards Philipp Kern signature.asc Description: Digital signature
Re: Proposal to add patches to netcfg (#682737)
Hi there! On Tue, 20 Nov 2012 12:42:14 +0100, Philipp Kern wrote: > * ba5ee42 Write out network-manager configuration files with mode 0600. [...] > * 4e1dbbd finish-install: Do not set /etc/network/interfaces to mode 0600. [...] > So we need libuuid only on non-Linux platforms (I hope it still compiled > there). As discussed the mode of both n-m config files and /e/n/i was > adjusted to be private to root. If I read the Git commits above correctly, there is a discrepancy with the latest sentence (which by itself is an important change WRT the previous behavior in a "normal" Debian installation). Thx, bye, Gismo / Luca pgpavbcr3da0L.pgp Description: PGP signature
Re: Proposal to add patches to netcfg (#682737)
On 20.11.2012 14:13, Luca Capello wrote: > Hi there! > > On Tue, 20 Nov 2012 12:42:14 +0100, Philipp Kern wrote: >> * ba5ee42 Write out network-manager configuration files with mode 0600. > [...] NM config files need to be 0600, so this is correct. >> * 4e1dbbd finish-install: Do not set /etc/network/interfaces to mode 0600. > [...] In the test install that I did with the mini.iso provided by Phil, /e/n/i was mode 0644. But since I did a desktop installation, I assume this file was created / modified by finish-install? >> So we need libuuid only on non-Linux platforms (I hope it still compiled >> there). This is a bit weird. The UUID is only required really on Linux plattforms, since NM is Linux-only. I would expect the NM specific code is ifdeffed on those plattforms so a libuuid dependency is not needed. Is this not the case? As discussed the mode of both n-m config files and /e/n/i was >> adjusted to be private to root. > > If I read the Git commits above correctly, there is a discrepancy with > the latest sentence (which by itself is an important change WRT the > previous behavior in a "normal" Debian installation). Agreed, clarification here would be good. Especially since the testing showed, that they are both not 0600. Michael -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth? signature.asc Description: OpenPGP digital signature
Re: Proposal to add patches to netcfg (#682737)
Luca, am Tue, Nov 20, 2012 at 02:13:37PM +0100 hast du folgendes geschrieben: > On Tue, 20 Nov 2012 12:42:14 +0100, Philipp Kern wrote: > > * ba5ee42 Write out network-manager configuration files with mode 0600. > [...] > > * 4e1dbbd finish-install: Do not set /etc/network/interfaces to mode 0600. > [...] > > So we need libuuid only on non-Linux platforms (I hope it still compiled > > there). As discussed the mode of both n-m config files and /e/n/i was > > adjusted to be private to root. > If I read the Git commits above correctly, there is a discrepancy with > the latest sentence (which by itself is an important change WRT the > previous behavior in a "normal" Debian installation). yeah, I screwed that up in the mail only. /e/n/i is supposed to be 0644, sorry about stating it incorrectly. We discussed that there's no need to have it 0600 in private and somehow I wasn't awake enough tomorrow morning (I added d-boot at the very last minute). Kind regards Philipp Kern signature.asc Description: Digital signature
Bug#690586: installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2
Am 20.11.2012 um 08:49 schrieb Andreas B. Mundt: > Find attached two netconsole logs that look more helpful: "unknown i2c > port". Interesting news: The machine does not lock up completely, > the log continues and after a few minutes I switched off the machine > by pressing the power button for a short time. Could you please try a 3.3 kernel from snapshot.debian.org¹? There were various i2c related nouveau commits in that version, so maybe it would be sufficient to backport them. Cheers, Sven ¹ http://snapshot.debian.org/archive/debian/20120322T225438Z/pool/main/l/linux-2.6/linux-image-3.3.0-trunk-amd64_3.3-1%7Eexperimental.1_amd64.deb -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87vcd06wlk@turtle.gmx.de
Bug#690586: installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2
Hi Sven, On Tue, Nov 20, 2012 at 05:57:27PM +0100, Sven Joachim wrote: [...] > Could you please try a 3.3 kernel from snapshot.debian.org¹? There were > various i2c related nouveau commits in that version, so maybe it would > be sufficient to backport them. > ¹ > http://snapshot.debian.org/archive/debian/20120322T225438Z/pool/main/l/linux-2.6/linux-image-3.3.0-trunk-amd64_3.3-1%7Eexperimental.1_amd64.deb This kernel works fine without any special handling, just installing and rebooting is sufficient. $ uname -a Linux lmz 3.3.0-trunk-amd64 #1 SMP Thu Mar 22 18:02:10 UTC 2012 x86_64 GNU/Linux Thanks! Andi -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121120182922.GA4561@lmz
Processed: Re: Bug#690586: installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2
Processing control commands: > severity -1 important Bug #690586 [installation-reports] installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2 Severity set to 'important' from 'normal' > reassign -1 src:linux Bug #690586 [installation-reports] installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2 Bug reassigned from package 'installation-reports' to 'src:linux'. Ignoring request to alter found versions of bug #690586 to the same values previously set Ignoring request to alter fixed versions of bug #690586 to the same values previously set > found -1 3.2.23-1 Bug #690586 [src:linux] installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2 Marked as found in versions linux/3.2.23-1. > fixed -1 3.3-1~experimental.1 Bug #690586 [src:linux] installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2 The source 'linux' and version '3.3-1~experimental.1' do not appear to match any binary packages Marked as fixed in versions linux/3.3-1~experimental.1. -- 690586: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=690586 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/handler.s.b690586.13534382185648.transcr...@bugs.debian.org
Bug#690586: installation-reports: NVIDIA GF108 [Quadro 1000M] 10de:0dfa (rev a1) X fails with nouveau/linux-3.2
Control: severity -1 important Control: reassign -1 src:linux Control: found -1 3.2.23-1 Control: fixed -1 3.3-1~experimental.1 Am 20.11.2012 um 19:29 schrieb Andreas B. Mundt: > On Tue, Nov 20, 2012 at 05:57:27PM +0100, Sven Joachim wrote: > [...] >> Could you please try a 3.3 kernel from snapshot.debian.org¹? There were >> various i2c related nouveau commits in that version, so maybe it would >> be sufficient to backport them. > >> ¹ >> http://snapshot.debian.org/archive/debian/20120322T225438Z/pool/main/l/linux-2.6/linux-image-3.3.0-trunk-amd64_3.3-1%7Eexperimental.1_amd64.deb > > This kernel works fine without any special handling, just installing > and rebooting is sufficient. > > $ uname -a > Linux lmz 3.3.0-trunk-amd64 #1 SMP Thu Mar 22 18:02:10 UTC 2012 x86_64 > GNU/Linux Thank you. I'm reassigning the bug to the Linux kernel and hopefully mark it as fixed in 3.3 and later. Now the question is which commits would need to be backported to the wheezy kernel. Some candidates: , | $ git log --grep="nouveau.*i2c" --pretty=oneline v3.2..v3.3 | 525895ba388c949aa906f26e3ec5cb1ab041f56b drm/nouveau/gem: fix fence_sync race / oops | 5d56fe5fd794a98c4f446f8665fd06b82e93ff64 Merge branch 'drm-nouveau-next' of git://anongit.freedeskto | f553b79c03f0dbd52f6f03abe8233a2bef8cbd0d drm/nouveau/i2c: handle bit-banging ourselves | 9e3b6b99075a01ce47379eee74365aece14242f3 drm/nouveau/i2c: fix debug message | 2bdb06e3cff066c546fb41152bc582a5ec73e899 drm/nouveau/i2c: tidy up bit-bang helpers, also fixing nv50 | 486a45c2a6c19b159602d044ab601a92cd81f524 drm/nouveau/i2c: do parsing of i2c-related vbios info in no | 0f8067c7054d22f240fca376e01430eecdc112df drm/nouveau/bios: fold fixup_legacy_i2c ` But that's beyond my skills :-(, so I'm passing the ball to the Debian kernel team. Cheers, Sven -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/87mwyc5c73@turtle.gmx.de
Bug#693839: debian-installer: kernel install fails on armel buffalo linkstation pro, missing uboot-mkimage
Package: debian-installer Version: debian armel beta3 installer on linkstation pro Severity: grave Tags: d-i Justification: renders package unusable Dear Maintainer, Clean installing Debian 7 with Beta3 netinstall on Linkstation Pro/Live fails kernel install. Started by copying uImage and initrd of installer to first disk partition and starting linkstation. Complete format of disk with partman in installer. Errorlog shows missing mkimage program at failed kernal install stage. Going into shell, chroot into /target, installing uboot-mkimage and restarting kernel-install task fixes the problem. This is maybe related to uboot-mkimage beeing transitional package (pulls u-boot-tools). -- System Information: Debian Release: wheezy/sid APT prefers testing APT policy: (500, 'testing') Architecture: armel (armv5tel) Kernel: Linux 3.2.0-4-orion5x Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8) Shell: /bin/sh linked to /bin/dash Log Fail: Nov 20 20:47:40 in-target: update-initramfs: Generating /boot/initrd.img-3.2.0-4-orion5x Nov 20 20:48:16 in-target: flash-kernel: installing version 3.2.0-4-orion5x Nov 20 20:48:19 in-target: Generating kernel u-boot image... /usr/sbin/flash-kernel: 232: /usr/sbin/flash-kernel: mkimage: not found Nov 20 20:48:19 flash-kernel-installer: error: flash-kernel failed Nov 20 20:48:19 main-menu[1243]: WARNING **: Configuring 'flash-kernel-installer' failed with error code 1 Nov 20 20:47:40 in-target: update-initramfs: Generating /boot/initrd.img-3.2.0-4-orion5x Nov 20 20:48:16 in-target: flash-kernel: installing version 3.2.0-4-orion5x Nov 20 20:48:19 in-target: Generating kernel u-boot image... /usr/sbin/flash-kernel: 232: /usr/sbin/flash-kernel: mkimage: not found Nov 20 20:48:19 flash-kernel-installer: error: flash-kernel failed Nov 20 20:48:19 main-menu[1243]: WARNING **: Configuring 'flash-kernel-installer' failed with error code 1 Nov 20 20:48:19 main-menu[1243]: WARNING **: Menu item 'flash-kernel-installer' failed. Log after uboot-mkimage install in target: Nov 20 21:03:09 main-menu[1243]: DEBUG: resolver (libgcc1): package doesn't exist (ignored) Nov 20 21:03:12 main-menu[1243]: INFO: Menu item 'flash-kernel-installer' selected Nov 20 21:03:22 in-target: Reading package lists... Nov 20 21:03:22 in-target: Nov 20 21:03:22 in-target: Building dependency tree... Nov 20 21:03:25 in-target: Nov 20 21:03:25 in-target: Reading state information... Nov 20 21:03:25 in-target: Nov 20 21:03:27 in-target: flash-kernel is already the newest version. Nov 20 21:03:27 in-target: 0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded. Nov 20 21:03:30 in-target: update-initramfs: Generating /boot/initrd.img-3.2.0-4-orion5x Nov 20 21:04:05 in-target: flash-kernel: installing version 3.2.0-4-orion5x Nov 20 21:04:08 in-target: Generating kernel u-boot image... Nov 20 21:04:08 in-target: done. Nov 20 21:04:08 in-target: Installing new uImage.buffalo. Nov 20 21:04:08 in-target: Generating initramfs u-boot image... Nov 20 21:04:09 in-target: done. Nov 20 21:04:09 in-target: Installing new initrd.buffalo. Nov 20 21:04:09 main-menu[1243]: INFO: Restoring default debconf priority 'high' -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121120213138.2147.73318.report...@debian.example.org
Bug#693485: Re[2]: Bug#693485: keyboard-configuration: XKBLAYOUT, XKBOPTIONS values manually set in /etc/default/keyboard lost on upgrade
Anton Zinoviev: > 3. Whether this bug is reproducible or console-setup no longer > overwrites your manual settings. I've noticed this behavior on 2 Debian installations; it occured on package upgrade (1.82 -> 1.87) only and doesn't happen on every reboot. I'm not sure whether it's caused by unconsolidated configs handling (?) by console-setup package install scripts or whatever. > 1. The original values of XKBLAYOUT and XKBOPTIONS. > > 2. The changed values of XKBLAYOUT and XKBOPTIONS. > 1) Machine #1: --- keyboard.old +++ keyboard.new @@ -3,8 +3,8 @@ # Consult the keyboard(5) manual page. XKBMODEL="acer_laptop" -XKBLAYOUT="us,ua,ru" +XKBLAYOUT="" XKBVARIANT="" -XKBOPTIONS="grp:alt_shift_toggle" +XKBOPTIONS="" BACKSPACE="guess" 2) Machine #2: --- keyboard.old +++ keyboard.new @@ -5,6 +5,6 @@ XKBMODEL="genius" XKBLAYOUT="us,ua,ru" XKBVARIANT="" -XKBOPTIONS="grp:alt_shift_toggle" +XKBOPTIONS="grp:alt_shift_toggle,grp_led:scroll" BACKSPACE="guess"
Processed: Re: Bug#693839: debian-installer: kernel install fails on armel buffalo linkstation pro, missing uboot-mkimage
Processing control commands: > reassign -1 flash-kernel 3.3 Bug #693839 [debian-installer] debian-installer: kernel install fails on armel buffalo linkstation pro, missing uboot-mkimage Bug reassigned from package 'debian-installer' to 'flash-kernel'. Ignoring request to alter found versions of bug #693839 to the same values previously set Ignoring request to alter fixed versions of bug #693839 to the same values previously set Bug #693839 [flash-kernel] debian-installer: kernel install fails on armel buffalo linkstation pro, missing uboot-mkimage Marked as found in versions flash-kernel/3.3. > severity -1 serious Bug #693839 [flash-kernel] debian-installer: kernel install fails on armel buffalo linkstation pro, missing uboot-mkimage Severity set to 'serious' from 'grave' -- 693839: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=693839 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/handler.s.b693839.135345328411902.transcr...@bugs.debian.org
Bug#693839: debian-installer: kernel install fails on armel buffalo linkstation pro, missing uboot-mkimage
Control: reassign -1 flash-kernel 3.3 Control: severity -1 serious Suschman (20/11/2012): > Package: debian-installer > Version: debian armel beta3 installer on linkstation pro > Severity: grave > Tags: d-i > Justification: renders package unusable > > Dear Maintainer, > > Clean installing Debian 7 with Beta3 netinstall on Linkstation > Pro/Live fails kernel install. Started by copying uImage and initrd > of installer to first disk partition and starting linkstation. > Complete format of disk with partman in installer. > > Errorlog shows missing mkimage program at failed kernal install > stage. Going into shell, chroot into /target, installing > uboot-mkimage and restarting kernel-install task fixes the problem. > This is maybe related to uboot-mkimage beeing transitional package > (pulls u-boot-tools). Hi, and thanks for your report. u-boot-tools is only in “Suggests” for flash-kernel, while your situation suggests it might be better to have it in “Recommends”. What do others think? Mraw, KiBi. signature.asc Description: Digital signature
Re: Bug#651720: new ZFS installs completely broken in Wheezy/Sid
Hi Steven, On 11/16/2012 10:42 PM, Steven Chamberlain wrote: > I noticed some relevant code in the grub-installer script that seems to > do that: > >> if [ "$bootfstype" = "zfs" ]; then >> # Required by update-grub on ZFS >> mkdir -p $ROOT/boot/zfs >> cp /boot/zfs/zpool.cache $ROOT/boot/zfs/ >> fi > > but of course that will only work where /boot itself resides on ZFS. > Maybe this needs to also match where $rootfstype is "zfs". I am not sure if checking for / being ZFS is good enough then. Wouldn't we need to check for any file system being on ZFS to create a zpool.cache file? Aside, your triaging sounds like a good explanation and like an easy fix. Thing is, my d-i fu is non-existing. Would anyone care to help? I've prepared a _completely untested_ patch which may fix the issue but I have no idea how to test it. Note, if you are right, this bug is filed against the wrong package, we may clone and reassign it to grub-installer then. -- with kind regards, Arno Töll IRC: daemonkeeper on Freenode/OFTC GnuPG Key-ID: 0x9D80F36D From b08feadb9604ba41a9a63f73edc15b81d42b6ef0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Arno=20T=C3=B6ll?= Date: Wed, 21 Nov 2012 00:15:35 +0100 Subject: [PATCH] Also create /boot/zfs/zpool.cache if the root fs is on ZFS --- grub-installer |6 -- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/grub-installer b/grub-installer index d709de0..f01eda1 100755 --- a/grub-installer +++ b/grub-installer @@ -221,8 +221,10 @@ rootfs=$(findfs /) bootfs=$(findfs /boot) [ -n "$bootfs" ] || bootfs="$rootfs" + bootfstype=$(findfstype /boot) -[ -n "$bootfstype" ] || bootfstype="$(findfstype /)" +rootfstype="$(findfstype /)" +[ -n "$bootfstype" ] || bootfstype=$rootfstype case $ARCH in powerpc/chrp|powerpc/chrp_rs6k|powerpc/chrp_ibm|powerpc/cell) @@ -851,7 +853,7 @@ fi db_progress STEP 1 db_progress INFO grub-installer/progress/step_config_loader -if [ "$bootfstype" = "zfs" ]; then +if [ "$bootfstype" = "zfs" ] || [ "$rootfstype" = "zfs" ] ; then # Required by update-grub on ZFS mkdir -p $ROOT/boot/zfs cp /boot/zfs/zpool.cache $ROOT/boot/zfs/ -- 1.7.10.4 signature.asc Description: OpenPGP digital signature
Bug#651720: Fwd: Re: Bug#651720: new ZFS installs completely broken in Wheezy/Sid
[Err, sorry sent the message everywhere but to the bug] Hi Steven, On 11/16/2012 10:42 PM, Steven Chamberlain wrote: > I noticed some relevant code in the grub-installer script that seems to > do that: > >> if [ "$bootfstype" = "zfs" ]; then >> # Required by update-grub on ZFS >> mkdir -p $ROOT/boot/zfs >> cp /boot/zfs/zpool.cache $ROOT/boot/zfs/ >> fi > > but of course that will only work where /boot itself resides on ZFS. > Maybe this needs to also match where $rootfstype is "zfs". I am not sure if checking for / being ZFS is good enough then. Wouldn't we need to check for any file system being on ZFS to create a zpool.cache file? Aside, your triaging sounds like a good explanation and like an easy fix. Thing is, my d-i fu is non-existing. Would anyone care to help? I've prepared a _completely untested_ patch which may fix the issue but I have no idea how to test it. Note, if you are right, this bug is filed against the wrong package, we may clone and reassign it to grub-installer then. From b08feadb9604ba41a9a63f73edc15b81d42b6ef0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Arno=20T=C3=B6ll?= Date: Wed, 21 Nov 2012 00:15:35 +0100 Subject: [PATCH] Also create /boot/zfs/zpool.cache if the root fs is on ZFS --- grub-installer |6 -- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/grub-installer b/grub-installer index d709de0..f01eda1 100755 --- a/grub-installer +++ b/grub-installer @@ -221,8 +221,10 @@ rootfs=$(findfs /) bootfs=$(findfs /boot) [ -n "$bootfs" ] || bootfs="$rootfs" + bootfstype=$(findfstype /boot) -[ -n "$bootfstype" ] || bootfstype="$(findfstype /)" +rootfstype="$(findfstype /)" +[ -n "$bootfstype" ] || bootfstype=$rootfstype case $ARCH in powerpc/chrp|powerpc/chrp_rs6k|powerpc/chrp_ibm|powerpc/cell) @@ -851,7 +853,7 @@ fi db_progress STEP 1 db_progress INFO grub-installer/progress/step_config_loader -if [ "$bootfstype" = "zfs" ]; then +if [ "$bootfstype" = "zfs" ] || [ "$rootfstype" = "zfs" ] ; then # Required by update-grub on ZFS mkdir -p $ROOT/boot/zfs cp /boot/zfs/zpool.cache $ROOT/boot/zfs/ -- 1.7.10.4 signature.asc Description: OpenPGP digital signature
Re: Bug#651720: new ZFS installs completely broken in Wheezy/Sid
Hi Arno! Sorry, I somehow didn't have 651...@bugs.debian.org copied on my previous mail so I'm inlining your mail for the benefit of the BTS: On 20/11/12 23:19, Arno Töll wrote: > I am not sure if checking for / being ZFS is good enough then. Wouldn't > we need to check for any file system being on ZFS to create a > zpool.cache file? At boot time, GRUB must provide kFreeBSD with a zpool.cache so that it can mount the root filesystem. After that I don't think it matters about other partitions being on ZFS, because they can be scanned for, it just may be a little slower first time around. If root is UFS and only /boot is ZFS, I'm not sure that the zpool.cache is still needed, but it can't hurt to preserve the existing behaviour for now and generate one anyway. > Aside, your triaging sounds like a good explanation and like an easy > fix. Thing is, my d-i fu is non-existing. Would anyone care to help? > I've prepared a _completely untested_ patch which may fix the issue but > I have no idea how to test it. Your patch looks perfect to me, but likewise I haven't figured out how to build+test it yet. > Note, if you are right, this bug is filed against the wrong package, we > may clone and reassign it to grub-installer then. I'm including Arno's patch, attached. Thanks, Regards, -- Steven Chamberlain ste...@pyro.eu.org From b08feadb9604ba41a9a63f73edc15b81d42b6ef0 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Arno=20T=C3=B6ll?= Date: Wed, 21 Nov 2012 00:15:35 +0100 Subject: [PATCH] Also create /boot/zfs/zpool.cache if the root fs is on ZFS --- grub-installer |6 -- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/grub-installer b/grub-installer index d709de0..f01eda1 100755 --- a/grub-installer +++ b/grub-installer @@ -221,8 +221,10 @@ rootfs=$(findfs /) bootfs=$(findfs /boot) [ -n "$bootfs" ] || bootfs="$rootfs" + bootfstype=$(findfstype /boot) -[ -n "$bootfstype" ] || bootfstype="$(findfstype /)" +rootfstype="$(findfstype /)" +[ -n "$bootfstype" ] || bootfstype=$rootfstype case $ARCH in powerpc/chrp|powerpc/chrp_rs6k|powerpc/chrp_ibm|powerpc/cell) @@ -851,7 +853,7 @@ fi db_progress STEP 1 db_progress INFO grub-installer/progress/step_config_loader -if [ "$bootfstype" = "zfs" ]; then +if [ "$bootfstype" = "zfs" ] || [ "$rootfstype" = "zfs" ] ; then # Required by update-grub on ZFS mkdir -p $ROOT/boot/zfs cp /boot/zfs/zpool.cache $ROOT/boot/zfs/ -- 1.7.10.4 signature.asc Description: OpenPGP digital signature
Processing of debootstrap_1.0.44_amd64.changes
debootstrap_1.0.44_amd64.changes uploaded successfully to localhost along with the files: debootstrap_1.0.44.dsc debootstrap_1.0.44.tar.gz debootstrap_1.0.44_all.deb debootstrap-udeb_1.0.44_all.udeb Greetings, Your Debian queue daemon (running on host franck.debian.org) -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e1taxlb-0007gt...@franck.debian.org
Bug#693718: marked as done (debootstrap: wrong 'for' in kldstat checks for freebsd)
Your message dated Tue, 20 Nov 2012 23:47:28 + with message-id and subject line Bug#693718: fixed in debootstrap 1.0.44 has caused the Debian Bug report #693718, regarding debootstrap: wrong 'for' in kldstat checks for freebsd to be marked as done. This means that you claim that the problem has been dealt with. If this is not the case it is now your responsibility to reopen the Bug report if necessary, and/or fix the problem forthwith. (NB: If you are a system administrator and have no idea what this message is talking about, this may indicate a serious mail system misconfiguration somewhere. Please contact ow...@bugs.debian.org immediately.) -- 693718: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=693718 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems --- Begin Message --- Package: debootstrap Version: 1.0.42 In the deboostrap scripts the string in sh: " for module in "linprocfs fdescfs tmpfs linsysfs"; do " it is interpreted as one string, so kldstat -m "$module" > /dev/null 2>&1 is always produce warning. Please remove double quotes for fix problem: --- /usr/local/sbin/debootstrap.bak 2012-11-19 19:44:00.529532488 +0400 +++ /usr/local/sbin/debootstrap 2012-11-19 19:48:17.165541277 +0400 @@ -437,7 +437,7 @@ fi if [ "$HOST_OS" = "kfreebsd" ] || [ "$HOST_OS" = "freebsd" ]; then - for module in "linprocfs fdescfs tmpfs linsysfs"; do + for module in linprocfs fdescfs tmpfs linsysfs; do kldstat -m "$module" > /dev/null 2>&1 || warning SANITYCHECK "Probably required module %s is not loaded" "$module" done fi --- End Message --- --- Begin Message --- Source: debootstrap Source-Version: 1.0.44 We believe that the bug you reported is fixed in the latest version of debootstrap, which is due to be installed in the Debian FTP archive. A summary of the changes between this version and the previous one is attached. Thank you for reporting the bug, which will now be closed. If you have further comments please address them to 693...@bugs.debian.org, and the maintainer will reopen the bug report if appropriate. Debian distribution maintenance software pp. Cyril Brulebois (supplier of updated debootstrap package) (This message was generated automatically at their request; if you believe that there is a problem with it please contact the archive administrators by mailing ftpmas...@debian.org) -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Format: 1.8 Date: Tue, 20 Nov 2012 23:55:53 +0100 Source: debootstrap Binary: debootstrap debootstrap-udeb Architecture: source all Version: 1.0.44 Distribution: unstable Urgency: low Maintainer: Debian Install System Team Changed-By: Cyril Brulebois Description: debootstrap - Bootstrap a basic Debian system debootstrap-udeb - Bootstrap the Debian system (udeb) Closes: 693718 Changes: debootstrap (1.0.44) unstable; urgency=low . * Remove double quotes to fix for loop on GNU/kFreeBSD, thanks to Oleg Ginzburg (Closes: #693718). Checksums-Sha1: aaf7983ec9a565922243c8eda1e12248e0aed5fb 1149 debootstrap_1.0.44.dsc 58f94407ee8edb9e6d537c8c5b89ce86c903d593 58467 debootstrap_1.0.44.tar.gz 8a95ad7905a0e6eba5408799d22233d1ab0f25f2 62676 debootstrap_1.0.44_all.deb 339e274c6283034555483280ce2b8d902d1095d9 18002 debootstrap-udeb_1.0.44_all.udeb Checksums-Sha256: 624520a5f2a872ecf42ea1f990a30c8cec59be5adc732b2084b0b92c03c55399 1149 debootstrap_1.0.44.dsc 3a620a6caff157b8d46955024451f100d129dd41a2fe3fd1734f662e59a2c00b 58467 debootstrap_1.0.44.tar.gz 4642d1f91da44e702055a8de608d78716d6fc9df488c6cdf0a316fdb5a25e472 62676 debootstrap_1.0.44_all.deb 7c326faf444b85a136a332a5e5760a
debootstrap_1.0.44_amd64.changes ACCEPTED into unstable
Accepted: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Format: 1.8 Date: Tue, 20 Nov 2012 23:55:53 +0100 Source: debootstrap Binary: debootstrap debootstrap-udeb Architecture: source all Version: 1.0.44 Distribution: unstable Urgency: low Maintainer: Debian Install System Team Changed-By: Cyril Brulebois Description: debootstrap - Bootstrap a basic Debian system debootstrap-udeb - Bootstrap the Debian system (udeb) Closes: 693718 Changes: debootstrap (1.0.44) unstable; urgency=low . * Remove double quotes to fix for loop on GNU/kFreeBSD, thanks to Oleg Ginzburg (Closes: #693718). Checksums-Sha1: aaf7983ec9a565922243c8eda1e12248e0aed5fb 1149 debootstrap_1.0.44.dsc 58f94407ee8edb9e6d537c8c5b89ce86c903d593 58467 debootstrap_1.0.44.tar.gz 8a95ad7905a0e6eba5408799d22233d1ab0f25f2 62676 debootstrap_1.0.44_all.deb 339e274c6283034555483280ce2b8d902d1095d9 18002 debootstrap-udeb_1.0.44_all.udeb Checksums-Sha256: 624520a5f2a872ecf42ea1f990a30c8cec59be5adc732b2084b0b92c03c55399 1149 debootstrap_1.0.44.dsc 3a620a6caff157b8d46955024451f100d129dd41a2fe3fd1734f662e59a2c00b 58467 debootstrap_1.0.44.tar.gz 4642d1f91da44e702055a8de608d78716d6fc9df488c6cdf0a316fdb5a25e472 62676 debootstrap_1.0.44_all.deb 7c326faf444b85a136a332a5e5760a67a6572e3f091c97e599b723a265205ee6 18002 debootstrap-udeb_1.0.44_all.udeb Files: dab11c018f71dca94c864f7bb0ab132f 1149 admin extra debootstrap_1.0.44.dsc 73d965abf163b31002f754e3a5322a52 58467 admin extra debootstrap_1.0.44.tar.gz 678511a33a9943c747c1a75b4d6f7d6f 62676 admin extra debootstrap_1.0.44_all.deb 69393ee88f494b4a745eac69943996a8 18002 debian-installer extra debootstrap-udeb_1.0.44_all.udeb -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.12 (GNU/Linux) iEYEARECAAYFAlCsEuMACgkQeGfVPHR5Nd1AZgCfZTW82me2ha/z+FKU6VmyB3Tm iKsAnirS8Sf0mmN970wOMqQrqqWykr9N =abSB -END PGP SIGNATURE- Thank you for your contribution to Debian. -- To UNSUBSCRIBE, email to debian-boot-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/e1taxxm-0001af...@franck.debian.org
Bug#651720: new ZFS installs completely broken in Wheezy/Sid
On Tue, 20 Nov 2012, Steven Chamberlain wrote: Hi Arno! Sorry, I somehow didn't have 651...@bugs.debian.org copied on my previous mail so I'm inlining your mail for the benefit of the BTS: On 20/11/12 23:19, Arno Töll wrote: I am not sure if checking for / being ZFS is good enough then. Wouldn't we need to check for any file system being on ZFS to create a zpool.cache file? At boot time, GRUB must provide kFreeBSD with a zpool.cache so that it can mount the root filesystem. After that I don't think it matters about other partitions being on ZFS, because they can be scanned for, it just may be a little slower first time around. The zpool.cache is no longer needed in upstream FreeBSD svn trunk, though this is not yet in any released version. So, there is probably a limit on the duration this bug will be relevant for. -Ben Kaduk
