Bug#337324: can reproduce the bug
I can't reproduce this bug. The first part was #334373 which is fixed now in the subversion trunk. The second bug (start-stop-daemon.REAL) does not occure with debootstrap 0.3.3. AFAIK the start-stop-daeom.REAL bug always was caused by a faulty debian mirror. Please try debootstrap 0.3.3 so we can close this bug soon. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#343338: discover-modprobe does not work on amd64
Package: discover Version: 2.0.7-2.1 Severity: normal Debian Release: 3.1 Architecture: amd64 Kernel: Linux 2.6.8-11-amd64-k8-smp #1 SMP Wed Jun 1 00:01:27 CEST 2005 x86_64 GNU/Linux Versions of packages discover depends on: ii libdiscover2 2.0.7-2.1 hardware identification library When running discover-modprobe on amd64, no modules will be loaded. Here's the debug output: suenner:~# sh -x /sbin/discover-modprobe + EX_USAGE=64 + sysconfdir=/etc + localstatedir=/var + discover=discover + '[' -x /usr/bin/discover-config ']' ++ discover-config --sysconfdir + sysconfdir=/etc ++ discover-config --localstatedir + localstatedir=/var + '[' -x /sbin/discover ']' + discover=/sbin/discover + conf=/etc/discover-modprobe.conf + crashdir=/var/lib/discover + crashfile=/var/lib/discover/crash + . /etc/discover-modprobe.conf ++ types=all + '[' -d /var/lib/discover ']' + nop= + verbose=false + getopts nv ch + shift 0 + false +++ uname -r ++ grep -E -v '^ *$' ++ /sbin/discover --data-path=linux/module/name --data-path=linux/module/options '--format=%s %s' --data-version=2.6.8-11-amd64-k8-smp all + module_details= + module_details_uniq= + module_details= + false + '[' -f /var/lib/discover/crash ']' + exit 0 The call of /sbin/discover . does not echo any modules. When omitting the "all" string at the end of line and calling the same command without it, everything works fine. This bug may be important or serious, since without hardware detection, the system may not boot up. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#343344: fai-mirror does not download packages from other arch
Package: fai-server Version: 2.9 Severity: minor When creating a partitial mirror with fai-mirror, packages are not downloaded if they does not exist on the architecture where fai-mirror is running. Example: on amd64 WARNING: These unknown packages are removed from the installation list: kernel-image-2.6-386 -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#343361: fai-mirror does not download packages defined by aptitude-r, bug in install_packages
Package: fai-server Version: 2.9 Tags: patch fai-mirror does not download pacakges defined by a PACKAGES aptitude-r rule. Change this line at the end of install_packages: < @commands = qw/taskinst aptitude install/; > @commands = qw/taskinst aptitude install aptitude-r/; -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#337271: patch can't work
I like to include this patch, but it can't work. The function getopt is not extended, so the new option -I will not be recognized. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#322157: removing backtick feature
I plan to remove the feature of backtick evaluation. IMO it not usefull any more. Any comments? Otherwise I will close this bug. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#340608: fai: nfsroot should not be in /usr
> On Fri, 25 Nov 2005 00:56:17 -0800, Steve Langasek <[EMAIL PROTECTED]> > said: > Could you elaborate on why you believe this is an FHS violation? Is the fai > nfsroot not shareable, or is it not read-only? (I would expect an nfsroot > image to be both...) The FAI nfsroot IS shareable and read only for all clients! I also think this is not an FHS violation. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#340608: fai: nfsroot should not be in /usr
> On Fri, 25 Nov 2005 13:41:18 +0100, Pierre THIERRY <[EMAIL PROTECTED]> > said: > read-write object the user can modify and update... It belongs either to > /var or /srv (the latter I prefer, as it is clearly data for a service > exposed by the system). My future plans are to move it to /srv, but the question is, if it's really a FHS violation. I'm also wondering, why there are currently no Debian packages using /srv. I hesitate to be the first package using /srv. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#340608: fai: nfsroot should not be in /usr
> On Fri, 25 Nov 2005 16:13:24 +0100, Pierre THIERRY <[EMAIL PROTECTED]> > said: > The problem is, strictly speaking, using /srv would not be policy > compliant, I think, because there is no mention of /srv in the currently > included FHS. Maybe you should just usr /var/lib/fai and just be > prepared to switch to /srv/fai when possible. I like to skip the move to /var/lib/fai, and wait until I can finally move to /srv. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#305861: You have to add boot into disk_config
If your boot partition is not the first partition, you'll have to add "boot" as an extra option in disk_config after the semicolon. FAI 2.8.1 will have an addition bug fix. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#309209: automatically edited files must not be conffiles
OK, it will prepare a fix for this bug ASAP. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#309209: patch contains CVS files
I will not use your patch, since it contains patches for the CVS files, which would corrupt my CVS tree. Instead I will apply these changes by hand and will also add some changes that are missing in your patch. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307838: unsecure temp file
Package: fai Version: 2.8.1 Severity: serious The script lib/updatebase creates unsecure temp file when called from task softupdate. This violates the policy. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#307632: not rc, not a security issues
> On Thu, 5 May 2005 16:02:33 -0400, Joey Hess <[EMAIL PROTECTED]> said: > severity 307632 normal > can be performed on unsafely created temp files. The program in question > is run during a fai install, before the system is multiuser, and so its > unsafe temp files cannot be created. Since fai 2.8 fai can also do softupdates (updateing the system, not installaing from scratch), that means you can run fai on an already installed system, which is running in multiuser mode. IHO this means the bug is a security issue. Currently I'm preparing a patch for this, and will upload a new version in a few days. Will you set the severity back to serious? -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#301179: fai: Early hooks (example: setup) are not executed.
I can't find any differences in the source code between 2.6.5 and 2.6.7 which could cause this error. Please try to detect the problem in more detail. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#313440: classes are not always evaluated during softupdate
When calling fai softupdate, fai uses the classes defined during the
installation. It does not call fai-class to execute scripts in
/fai/class to define new classes. This is only done when using fai -N
softupdate. So defining a class FAI_{$FAI_ACTION} with a scipt in
/fai/class will not work always, it willwork only if -N is used.
--
regards Thomas
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#313440: nice idea, but too early for defining this class
There may be some people that will set FAI_ACTION in a scripts /fai/*.var (this was used in LAST.var until fai 2.5.1), because they do not use DHCP or pxelinux.cfg for setting the action. Since *.var will be executed after defining classes, the patch will fail in these environments. We should discuss this problem again before applying the patch. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#320024: (no subject)
Package: fai Version: 2.8.4 Severity: important The script fai-cd does not work with sarge, since it has the string testing hardcoded in line 214. Just replace the string testing with stable. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#301342: (no subject)
Subject: does not find all .ko kernel drivers Package: mkinitrd-cd Version: 0.50 Severity: important Tags: patch *** Please type your report below this line *** Hi Rene, the new mkinitrd-cd version is fantastic, but it does not work at all for a 2.6 kernel, since it does not find all .ko kernel modules needed. The patch is attached. regards thomas pittermaennche[/usr/sbin]> diff -u mkinitrd-cd.orig mkinitrd-cd --- mkinitrd-cd.orig2005-03-06 19:05:23.0 +0100 +++ mkinitrd-cd 2005-03-25 09:56:06.528966352 +0100 @@ -124,10 +124,7 @@ if [ -d $modulepath/$n ] ; then for f in `find $modulepath/$n/ -name '*.o' -o -name '*.ko'`; do # copy only those that are also listed in scsiprobe.dat - mod=`expr $(basename $f) : '\(.*\).o'` - if [ -z "$mod" ]; then -mod=`expr $(basename $f) : '\(.*\).ko'` - fi + mod=`expr $(basename $f) : '\(.*\)\.[k]*o'` echo -ne "$mod\t" # cosmetics (another tab for short file names) if [ `expr "$mod" : '.*'` -lt 4 ]; then -- System Information: Debian Release: 3.1 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Kernel: Linux 2.6.9 Locale: LANG=C, LC_CTYPE=en_US (charmap=ISO-8859-1) Versions of packages mkinitrd-cd depends on: ii discover-data 2.2005.02.13-1 Data lists for Discover hardware d ii dosfstools2.10-1 Utilities to create and check MS-D ii gcc 4:3.3.5-1 The GNU C compiler ii libuclibc-dev 0.9.26-cvs20040816-5.1 A small implementation of the C li ii libuclibc00.9.26-cvs20040816-5.1 A small implementation of the C li ii make 3.80-9 The GNU version of the "make" util ii makedev 2.3.1-77 creates device files in /dev ii mtools3.9.9-2.1 Tools for manipulating MSDOS files ii syslinux 2.11-0.1 Bootloader for Linux/i386 using MS ii uclibc-toolchain 0.9.26-cvs20040816-5.1 A compiler wrapper for uClibc -- no debconf information -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#303565: make-fai-nfsroot is broken because is uses ext2resize
Package: fai Version: 2.7 Severity: grave make-fai-nfsroot is broken. It uses ext2resize, which is currently unavailable in testing. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#304179: missing packages in mirror when using fai-mirror
Package: fai Version: 2.8 Severity: normal Tags: patch A mirror created by fai-mirror does not contains the packages listed in $NFSROOT/var/tmp/packages.arch, even when giving the -a option. This is the patch: diff -r1.17 fai-mirror 186,187c186,187 < if [ -f $NFSROOT/var/tmp/var/tmp/packages.arch ]; then < plist=$(< $NFSROOT/var/tmp/base-pkgs.lis) --- > if [ -f $NFSROOT/var/tmp/packages.arch ]; then > plist=$(< $NFSROOT/var/tmp/packages.arch) -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#334333: fai-cd installs systems lacking a grub bootloader when using simple examples
install_packages should bepatched to download-only also package names ending with a minus when install_packages is called from fai-mirror. I will try to write a patch. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#334333: new patch for #334333
This is a propose for bug #334333. Please test it and send comments on it. [~]> diff fai-2.8.4/scripts/install_packages . 325c325,326 < unshift @orig, $_; --- > my $x = $opt_d ? $pack : $_; # when using -d, remove suffix -, so this > package will also be downloaded > unshift @orig, $x; -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#334333: fai-cd installs systems lacking a grub bootloader when using simple examples
> On Tue, 18 Oct 2005 09:53:57 +0200, Henning Glawe <[EMAIL PROTECTED]> > said: > packages should be only pulled onto the fai-cd if there is any "positive" > occurrence of a package... Your are right. So I have just to ignore any package name ending with a minus. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#334333: fai-cd installs systems lacking a grub bootloader when using simple examples
> On Tue, 18 Oct 2005 09:53:57 +0200, Henning Glawe <[EMAIL PROTECTED]> > said: > this is definitely not the right solution for the problem; there could be > just any reason to have a uninstalled in the package lists (especially, if > your $FAI for fai-cd and installations/softupdates is the same). > packages should be only pulled onto the fai-cd if there is any "positive" > occurrence of a package... What about this? [~]> diff fai-2.8.4/scripts/install_packages . 325c325,326 < unshift @orig, $_; --- > # when using -d, remove - suffix so this package will also be downloaded > unshift @orig, $_ unless ($opt_d && /-$/); -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#404187: fai-client: add NTFS support to setup_harddisks
Package: fai-client
Version: 3.1.3
Severity: wishlist
Tags: patch
setup_harddisks should support NTFS file systems.
Below is a patch that enables setup_harddisks to create a corret fstab
entry for a NTFS partition which can be preserved. This patch does not
add support for creating a NTFS file system with setup_harddisks.
--- setup_harddisks 2006-12-17 17:44:27.0 +0100
+++ setup_harddisks.new 2006-12-22 11:06:35.330107128 +0100
@@ -908,6 +908,7 @@
$device = $MountpointPart{$mountpoint};
$type = "ext2";
($MPOptions{$mountpoint} =~ /\b(dosfat16|winfat32)\b/i) && ($type =
"vfat");
+ ($MPOptions{$mountpoint} =~ /\b(ntfs)\b/i) && ($type = "ntfs");
($MPOptions{$mountpoint} =~ /\b(reiser)\b/i) && ($type = "reiserfs");
($MPOptions{$mountpoint} =~ /\b(xfs)\b/i) && ($type = "xfs");
($MPOptions{$mountpoint} =~ /\b(ext3)\b/i) && ($type = "ext3");
-- System Information:
Debian Release: 4.0
APT prefers testing
APT policy: (500, 'testing')
Architecture: i386 (i686)
Shell: /bin/sh linked to /bin/bash
Kernel: Linux 2.6.18-3-686
Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1)
Versions of packages fai-client depends on:
ii cfengine2 2.1.20-1 Tool for configuring and maintaini
ii file 4.17-5 Determines file type using "magic"
ii libapt-pkg-perl 0.1.20 Perl interface to libapt-pkg
ii perl 5.8.8-6.1 Larry Wall's Practical Extraction
fai-client recommends no packages.
-- no debconf information
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#405435: fai-doc: most Xorg output drivers are missing
Package: fai-doc Version: 3.1.3 Severity: normal Tags: patch xorg (1:7.1.0-9) moved the depends on xserver-xorg-(input/video)-all to a recommends. This causes FAI to install only very few video drivers. Following simple patch will fix it. Change a line in packages_config/XORG file to this: xorg xserver-xorg-video-all xserver-xorg-input-all -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#405433: fai-doc: xlockmore is not in testing any more
Package: fai-doc Version: 3.1.3 Severity: normal xlockmore is still used in the FAI class XORG, but this package was removed from testing. -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#407947: fai-server: fai-setup stops if chown -R of $loguserhome fails
Package: fai-server Version: 3.1.5 Severity: important fai-setup fails if the home directory of LOGUSER can't be written by root, for example, if it's an NFS mounted directory where the root account does not have write permissions. The chown -R command fails, and so the rest of the fai-setup script is not executed because set -e is used. -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) Versions of packages fai-server depends on: ii debootstrap 0.3.3.1Bootstrap a basic Debian system ii fai-client3.1.5 Fully Automatic Installation clien Versions of packages fai-server recommends: ii dhcp3-server 3.0.4-12 DHCP server for automatic IP addre ii fai-kernels 1.15 special kernels for FAI (Fully Aut ii nfs-kernel-server [nfs-server 1:1.0.10-4 Kernel NFS server support ii rsh-server0.17-13rsh servers. ii syslinux 3.31-1 Bootloader for Linux/i386 using MS ii tftpd-hpa [tftpd] 0.43-1 HPA's tftp server ii wget 1.10.2-2 retrieves files from the web -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#407950: fai-server: /etc/fai/NFSROOT contains package name, which do not exist any more
Package: fai-server Version: 3.1.5 Severity: minor When calling make-fai-nfsroot, packages from /etc/fai/NFSROOT are installed into the nfsroot. libnet-perl is not available any more in etch. It does not break the make-fai-nfsroot call, since the unknown package is detected and removed from the list of package that will be installed. -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) Versions of packages fai-server depends on: ii debootstrap 0.3.3.1Bootstrap a basic Debian system ii fai-client3.1.5 Fully Automatic Installation clien Versions of packages fai-server recommends: ii dhcp3-server 3.0.4-12 DHCP server for automatic IP addre ii fai-kernels 1.15 special kernels for FAI (Fully Aut ii nfs-kernel-server [nfs-server 1:1.0.10-4 Kernel NFS server support ii rsh-server0.17-13rsh servers. ii syslinux 3.31-1 Bootloader for Linux/i386 using MS ii tftpd-hpa [tftpd] 0.43-1 HPA's tftp server ii wget 1.10.2-2 retrieves files from the web -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#408153: fai-client: fai-class does define HOSTNAME any more
Package: fai-client Version: 3.1.5 Severity: grave Justification: renders package unusable The fix for #406125 breaks fai-class. Now, the HOSTNAME is not set correctly any more, so most classes are not defined. This will stop the installation when searching for a disk_config template. -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) Versions of packages fai-client depends on: ii cfengine2 2.1.20-1 Tool for configuring and maintaini ii file 4.17-5 Determines file type using "magic" ii libapt-pkg-perl 0.1.20 Perl interface to libapt-pkg ii perl 5.8.8-7Larry Wall's Practical Extraction fai-client recommends no packages. -- no debconf information -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#408247: fai-chboot: should read make-fai-nfsroot.conf to get NFSROOT
> On Wed, 24 Jan 2007 14:28:49 +0100, Raoul Borenius <[EMAIL PROTECTED]> > said: > fai-chboot still looks in fai.conf for $NFSROOT but this has moved > to make-fai-nfsroot.conf. Since NFSROOT is not defined in fai.conf, fai-chboot uses the default path of /srv/fai/nfsroot. So fai-chboot is not broken for most users. Puuh. I though this could be another grave bug, but it's not that critical. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#402644: My patch for this bug
I agree that we do not need the hash in the local log files. I wonder if it's a bug or a feature that we copy the hash (md5 by default) of the rootpw to the remote location. This fix may not be complete (depending on bug or feature that it's copied to remote), since fai-savelog copies from $LOGDIR when doing the remote copy and only $thislog/fai.log was cleaned up. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#402644: My patch for this bug
What do you think about making the the /var/log/fai directory read only for root and the group adm? IMO this would fix the security problem. - regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390367: --hostname and -u parsing dysfunctional
> On Sat, 9 Dec 2006 05:58:14 -0600, "Henning Sprang" <[EMAIL PROTECTED]> > said: > reopen 390367 Ich bin mir nicht sicher ob dieser bug wirklich erneut geoeffnet wurde. Auf jeden Fall ist die Subject Zeile nicht uebernommen worden. Vielleicht waere es besser gewesen, nicht den wishlist bug zu "reopen", sondern einfach einen normalen bug zu schreiben. Der wishlist bug wollte ja ein neues Feature haben, das wurde ja eingebaut. Nun hat dieses feature einen bug, also einfach einen neuen Bug dafuer schreiben. Das reopen ist IMO verwirrend. Ausserdem ist der bug jetzt ja nicht mehr wishlist, sondern normal, was das BTS aber nicht abbildet, es zeigt ihn weiter als wishlist an. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#406500: fai-doc: wrong path in fcopy example
Package: fai-doc Version: 3.1.4 Severity: normal examples/simple/scripts/FAISERVER/10-conffiles uses a wrong path for copying the sources.list file. It should be /etc/fai/apt/sources.list -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#400324: (no subject)
I'm pretty sure that this bug is solved. Please check again if this wrong path is really in the sources. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#400373: two problems with make-fai-bootfloppy
> On Sat, 25 Nov 2006 19:25:08 +0100, Holger Levsen <[EMAIL PROTECTED]> > said: > package: fai > version: 3.1.1 > severity: important Hallo Holger, bitte nicht zwei bugs in einem Bug report submitten. Das ist schlecht wenn wenn man nur einem Bug schliessen will. Ausserdem haette ich erwartet das du den Bug besser einstufst. > 2. Since fai-kernels 1.13 the kernel does not fit on 1.44mb floppies anymore, > so "-B" should be the default. I consider this important :) Das ist nicht important, da es nur ein kleineres Tool in FAI betrifft. Ausserdem ist dieses tool fuer die meisten Leute wohl nicht noetig, das sie PXE faehige Netzwerkkarten habe. "important: a major effect on the usability of a package" > P.S.: the IP address it's using as default or maybe only the output also looks > broken: Ja, das hatte ich auch schon bemerkt. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#400324: postinst info slightly inaccurate and confusing
> On Sat, 25 Nov 2006 12:40:28 +0100, Holger Levsen <[EMAIL PROTECTED]> > said: > And now I wonder why fai-server doesn't provide a /etc/fai/sources.list which > just works? :) It provides a sources.list which just works, but in most cases it's better to specify a Debian mirror that's nearer to you. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#400324: postinst info slightly inaccurate and confusing
> On Mon, 27 Nov 2006 20:17:28 +0100, Holger Levsen <[EMAIL PROTECTED]> > said: > On Monday 27 November 2006 20:06, you wrote: >> > And now I wonder why fai-server doesn't provide a >> > /etc/fai/sources.list which just works? :) >> It provides a sources.list > no it doesnt. /etc/fai/sources.list did not exist after a fresh install here. With FAI 3.X it's /etc/fai/apt/sources.list. The path has changed. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#389905: fai-chboot list wrong information
> On Wed, 25 Oct 2006 17:59:47 +0200, Thomas =?UTF-8?Q?P=C3=B6hnitzsch?=
> <[EMAIL PROTECTED]> said:
> +} elsif ($hex =~ /[0-9A-F]+/) {
> +my $hexstrip = (split /\./,$hex)[0]; # remove .disable if
necessary
> +$n = $hexstrip;
> +# hex to ip/subnet address
> +while ( length $n ) {
> +$host = sprintf( "%s%d" ,$host?"$host.":$host,
> +(hex substr $n,0,1,"" ) * 16 + (hex substr
$n,0,1,""));
> +}
> +if ( $host and ( length($hexstrip) < 8) ) {
> +$host = "Subnet: $host/". length($hexstrip)*4
Das verstehe ich nicht. Warum eine while schleife ueber $n? Oder soll
das nur wie ein if $n > 0 funktionieren? IMO macht die schleife aber
bei jedem Durschlauf das gleich wenn $n > 1 ist.
> - $srcfile = "$pxedir/$srchost";
> + $srcfile = "$pxedir/$srchost.tmpl";
Verstehe ich auch nicht.
Die man page ist schon gepatched. Leider ist dein patch sehr gross das
macht es schwer alles einzubauen. Und mein urspruenglicher bug ist
damit noch nicht gefixt, aber ich habe dafuer jetzt einen patch.
--
regards Thomas
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#358967: Remove fai-statoverride
> On Sun, 29 Oct 2006 23:43:17 +0100, Michael Tautschnig <[EMAIL > PROTECTED]> said: > What is fai-statoverride good for at all? A grep in the FAI sources did not turn > up any uses of the command, thus I suggest removing the script entirely. It's like dpkg-divert but changes things in the chroot of the new system and can handle a list of files. It's not used in the simples examples, but I use it this way: fai-statoverride root plugdev 4755 /usr/bin/pmount /usr/bin/pumount /usr/bin/wodim -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#389905: fai-chboot list wrong information
>>>>> On Wed, 1 Nov 2006 09:36:24 +0100, Thomas =?UTF-8?Q?P=C3=B6hnitzsch?=
>>>>> <[EMAIL PROTECTED]> said:
> On Sun, Oct 29, 2006 at 03:52:10PM +0100, Thomas Lange wrote:
>>> +while ( length $n ) {
>>> +$host = sprintf( "%s%d" ,$host?"$host.":$host,
>>> +(hex substr $n,0,1,"" ) * 16 + (hex substr
$n,0,1,""));
>>> +}
> Solange in $n noch etwas drin ist, verarbeite bis zu 2 Zeichen vom
> Anfang von $n. Das ist doch der klassische Trivialfall einer
> while-Schleife, und es funktioniert sogar. Ich verstehe daher nicht, was
> Dir daran nicht gefällt.
Sorry, mein Fehler. Ich habe nicht gesehen wo in der Schleife $n
veraendert wird. Jetzt habe ich mir von substr nochmals die genau
Funktion angesehen und jetzt sehe ich das $n innerhalb der schleife
auch veraendert wird.
> Soso, na das wundert mich jetzt aber, denn bei mir hat er den Bug
> behoben. Nur mal so wild aus dem Patch zitiert:
> +if (/\.disable$/) {
> +if (-f (split /\./)[0]) {
> +warn "Skipping $_ as disabled host because also enabled.\n"
if $verbose;
> +} else {
Ich werde es so machen. Deine Version geht wahrscheinlich nur wenn man
sich in dem pxelinux.cfg Verzeichnis befindet.
$f=(split (/\./,$_))[0];
if (-f "$pxedir/$f") {
> Es mag sein, daß Dir das nicht gefällt, aber behoben ist es definitiv,
> wie Du auch aus meinen Beispielen erkennen solltest.
Es geht nicht darum das mir etwas nicht gefaellt. Nein ich versuche
nur alles zu verstehen.
> Soll ich den Patch nochmal zerlegen? Warum sich das schwer einbauen läßt
> verstehe ich leider nicht, da ich den direkt gegen den aktuellen
> SVN-Head gebaut hatte, so daß ein einfaches "patch" ausreichen sollte.
Es liegt einfach daran das viele auf einmal schwerer zu versteh ist
als kleine Haeppchen. Aber jetzt ist das OK, ich werde alles nachher
einbauen. Viele Dank nochmals fuer die schnelle Antwort. Ist ja nicht
selbstverstaendlich das du so flink antwortest. Ich kenne da ganz
andere Intervalle von anderen :-)
--
regards Thomas
Bug#390346: (no subject)
Would it be OK to check if the path for dirinstall is an absolute path? If not, fai could print an error message. Or should fai convert a relative path to an absolute one? -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#361524: (no subject)
We have to check every variable in boot.log if it's needed during softupdates. fai action softupdates does not create or read boot.log, so those variables are not set. But maybe it's enough to document that LOGSERVER has to be defined in fai.conf if you want to use softupdate after the first installation. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390346: (no subject)
IMO the solution using abspath=$(cd $dir;pwd) to get the absolute path of dir does not work. When calling "fai dirinstall relpath" the user will specify a relative path that does not exist yet. So the cd will not work. Also using readlink -f will not help when the relative path does not exist. IMO it's better to force the user to specify an absolute path and to print a good error message in case he specifies a relative path. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390346: (no subject)
> On Thu, 2 Nov 2006 16:49:36 +0100, Holger Levsen <[EMAIL PROTECTED]> said: > If the directory doesn't exist, create it. Doesnt matter if the path is > relative or absolute. Sure this directory will be created. But it does matter in this case if it's relative or absolute. That's the bug. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#361524: (no subject)
Since FAI 3 the variable $LOGSERVER is not defined any more in fai.conf. You must define it in your config space in a *.var file. If you do not define it, fai-savelog has the fallback to use the variable $SERVER which is only defined during an installation (by get-boot-info) not during softupdate. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#289271: In addition, I believe that fai-setup:add_export_line() is broken
> On Sat, 08 Jan 2005 14:48:45 +0200, Shaul Karl <[EMAIL PROTECTED]> said: > The following patch should fix what looks to me broken: Does it only look broken, or do you have any p[roblems? > -[ -f /etc/exports ] && grep -q "^$pattern[[:space:]]" /etc/exports && return > -echo "$pattern $options" >> /etc/exports > +[ -f /etc/exports ] && ! grep -q "^$pattern[[:space:]]" /etc/exports && echo "$pattern $options" >> /etc/exports > +return > } > This patch: > 1. puts the return in a separate line. > 2. Negates grep's return value before using it in the list. I don't understand why this additional patch is needed. I never heard of any problems with these lines. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#290371: fai: LOGREMOTEDIR would be helpful even with rcp/scp
What about this in fai-savelog:
save_log_remote_shell() {
-local thislog=$HOSTNAME/$FAI_ACTION-$FAI_RUNDATE
+local thislog=$LOGREMOTEDIR/$HOSTNAME/$FAI_ACTION-$FAI_RUNDATE
Would that be ok for you?
--
regards Thomas
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#297550: make-fai-bootfloppy: 1.44M too small for kernel 2.6.10
Hi Richard, I can;t see the typo. Please specify in more detail. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#265444: FAI-CD available
The first version of FAI-CD is now available. Have a look at http://www.informatik.uni-koeln.de/fai/fai-cd/ -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#296932: fai-kernels: serial console doesn't work with 2.6 kernel
Since the 2.6 kernel is growing large, I had to remove some unused drivers, which are only rarely used. Otherwise this kernel will not fit onto a floppy disk. Should we explain it in more detail in the documentation? Do you like to write a patch for this? -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#409863: (no subject)
I would be very happy if the new memtest86+ package (containing version 1.70) would make it into etch. Please send a mail to debian-release@lists.debian.org to ask for unblocking the new version after uploading. This is a new upstream version but this should not matter. metest86+.bin can't can't break the linux system at all, since it's running without starting a kernel. The new version supports newer hardware so it's very usefull. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#410084: fai-doc: shell scripts do not work when /bin/sh is dash
Package: fai-doc Version: 3.1.6 Severity: important Most configuration scripts of the simple examples are shell scripts (with /bin/sh) that fail if /bin/sh is dash (as in ubuntu). Since they use bash only functions they should be forced to use /bin/bash. -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.18-3-686 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#413504: fai-client: setup_harddisks fails on creating reiserfs when not using locale en or C
Package: fai-client Version: 3.1.7 Severity: normal *** Please type your report below this line *** setup_harddisks calls mkreiserfs without using -q. So it needs to feed the answer for the question from mkreiserfs. This is done by echo 'y' | mreiserfs which does not work when the locale is set to non english questions/answers. More details http://www.uni-koeln.de/bin2/maillist/linux-fai/20070223.143038/174360 -- System Information: Debian Release: 4.0 APT prefers testing APT policy: (500, 'testing') Architecture: i386 (i686) Shell: /bin/sh linked to /bin/bash Kernel: Linux 2.6.20-486 Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) Versions of packages fai-client depends on: ii cfengine2 2.1.20-1 Tool for configuring and maintaini ii file 4.17-5 Determines file type using "magic" ii libapt-pkg-perl 0.1.20 Perl interface to libapt-pkg ii perl 5.8.8-7Larry Wall's Practical Extraction fai-client recommends no packages. -- no debconf information -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#412471: (no subject)
Please check if adding the next-server option fixes your problem. I'm running dhcp-server version 3 without any problems. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#413466: (no subject)
A non root user can call fai-mirror. You only need to add /usr/sbin in your PATH variable. Please check if this fixes your problem. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#413466: (no subject)
> On Wed, 7 Mar 2007 15:46:07 +0100, Holger Levsen <[EMAIL PROTECTED]> said: > If that fixes the problem, the documentation should > say "call /usr/sbin/fai-mirror" instead of "call fai-mirror" :-) No. It's /usr/sbin/install_packages. fai-mirror calls install_packages and this fails if /usr/sbin is not included in the PATH variable. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#416616: fai-client: fai-do-scripts confused by new file output: POSIX shell script
> On Thu, 29 Mar 2007 11:45:13 +0200, Federico Di Gregorio <[EMAIL > PROTECTED]> said: > Package: fai-client > Version: 3.1.8 > Severity: grave > Justification: renders package unusable > After upgrading to the last file package the file command returns Which version of the file package are you using? -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#416616: fai-client: fai-do-scripts confused by new file output: POSIX shell script
>>>>> On Thu, 29 Mar 2007 14:37:28 +0200, Federico Di Gregorio <[EMAIL >>>>> PROTECTED]> said: > Il giorno gio, 29/03/2007 alle 13.39 +0200, Thomas Lange ha scritto: >> >>>>> On Thu, 29 Mar 2007 11:45:13 +0200, Federico Di Gregorio <[EMAIL PROTECTED]> said: >> >> > Package: fai-client >> > Version: 3.1.8 >> > Severity: grave >> > Justification: renders package unusable >> >> > After upgrading to the last file package the file command returns >> >> Which version of the file package are you using? > 4.20-3 on the amd64 arch. Testing/Etch includes file package version 4.17-5. This works without problems. So I think this bug is only in unstable. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#416616: fai-client: fai-do-scripts confused by new file output: POSIX shell script
It was also suggested to use file -i in the future. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#411541: fai: nfs-user-server bug forces root squash and breaks ssh pub key access in fai
I'm wondering why the severity is important. AFAIK only saving the log files via ssh will not work, but not the whole installation is broken. Please provide additional information if more than this part is broken when using the nfs-user-server. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#412471: fai doesn't work with dhcp3-server
I heared that dhcp3 needs the next-server option. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#373587: reassign
I also think we can close the bug. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#385613: fai-kernels: Missing ethernet driver sk98lin on 2.6.16-fai, Marvell gigabit cards don't work
sk98lin is enabled, but maybe the two others are missing for your hardware. # CONFIG_SKGE is not set # CONFIG_SKY2 is not set CONFIG_SK98LIN=y -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#381494: (no subject)
Yes, IMO it's also usefull on amd64. I think all (or nearly all) network drivers that we use on I386 are also useful on amd64. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390447: also run task confdir with dirinstall
Variables from *.var are defined for dirinst since task_defvar is called before task_dirinst. Have a look at http://www.informatik.uni-koeln.de/fai/download/dirinst-log/fai.log -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390508: provide updated documentation (new features and changes of version 3.0 are not documented)
> Thomas, you could help others to help > you, by providing a list (in the wiki?) of what parts are missing. All user visible changes are described in NEWS.Debian. You could start using this list. I think I'm a little bit blind in doing the review of my own documentation. So I hope to get some detailed infos which parts should be improved or which information is missing or not well described. Patches are always welcome. Please do not write a huge bug patch for the documentation, but write small patches for each section or paragraph. Small patches are better to read an include into the documentation. If someone likes to help improving the documentation, please contact me before so we can coordinate this work. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390508: provide updated documentation (new features and changes of version 3.0 are not documented)
> Thomas, you could help others to help > you, by providing a list (in the wiki?) of what parts are missing. All user visible changes are described in NEWS.Debian. You could start using this list. I think I'm a little bit blind in doing the review of my own documentation. So I hope to get some detailed infos which parts should be improved or which information is missing or not well described. Patches are always welcome. Please do not write a huge bug patch for the documentation, but write small patches for each section or paragraph. Small patches are better to read an include into the documentation. If someone likes to help improving the documentation, please contact me before so we can coordinate this work. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390504: FAI_CONFIGDIR is ignored
IMO this bug is of severity normal. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390447: also run task confdir with dirinstall
> On Mon, 02 Oct 2006 14:32:52 +0200, Henning Sprang <[EMAIL PROTECTED]> > said: > Hmm. Probably I mixed that with the hostname problem (so, .var > wasn't executed, as well as .var for classes that i usually > define in a class/ file when doing netinstalls. If you mixed this up, please close this bug. I think adding a -h hostname option is an good idea. But we have a sparate wishlist bug for this. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390680: prevent concurrent fai runs
> On Mon, 2 Oct 2006 16:35:27 +0200, Holger Levsen <[EMAIL PROTECTED]> said:
> it would be great, if /usr/sbin/fai would check, if there is another
> fai-process running and if so, exit with an error-msg and an error.
Currently this is checked but not for softupdates
in bin/fai:
stamp=/var/run/fai/FAI_INSTALLATION_IN_PROGRESS
.
.
.
[ -f "$stamp" ] && {
echo "$0 already running, aborting"
exit 1
}
But fai softupdates uses another stamp file, that's the reason why it does
not work. from lib/subroutines:
>local stamp=/var/run/fai/fai_softupdate_is_running
Will be fixed soon.
--
regards Thomas
Bug#390944: Link to $FAI already exists
> On Tue, 3 Oct 2006 23:45:50 +0200, Michael Tautschnig <[EMAIL PROTECTED]> > said: > Updating SVN in /var/lib/fai/config > ln: creating symbolic link `/var/run/fai/current_config/config' to `/var/lib/fai/config': File exists So this link already points to the correct directory before calling ln? Or does it point to a different location? -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#384624: bug affects make-fai-bootfloppy
This bug does not affect make-fai-nfsroot but make-fai-bootfloppy. But I can confirm this bug. It will be fixed soon. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390345: (no subject)
Action names can be nearly any name. If this action is not one of the default actions, FAI tries to find a hook with the name of the action. Have a look at subroutines: *) if [ -f $FAI/hooks/$FAI_ACTION ]; then echo "Calling user defined action: $FAI_ACTION" $FAI/hooks/$FAI_ACTION else echo "ERROR: User defined action $FAI/hooks/$FAI_ACTION not found." -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390343: (no subject)
AFAIK the variable NFSROOT is only needed during dirinst when extracting the base.tgz from the nfsroot. I'll add a test which checks the existing of the base.tgz file. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390345: (no subject)
I propose to _change_ the action name from dirinst to dirinstall. So dirinst will not be possible any more. Since dirinst is pretty new, I have no problem to rename this action. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390346: (no subject)
Currently I do not remobmer which function returns the canonical path. How can I convert a relativ path to an absolute path? -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390345: (no subject)
> On Sat, 07 Oct 2006 16:51:26 +0200, Henning Sprang <[EMAIL PROTECTED]> > said: > Hmm, somehow I don't like the idea mixing up hooks and user defined > actions. Whare is this actually used? > When calling fai on the command line? yes, but also when doing network boot. So you can always define FAI_ACTION=myaction and if /srv/fai/config/hooks/myaction existst, it will be executed. > In my understanding, the definition of a hook is something that is run > _before_ an action with the same name, depending on a class that matches > the part in .. > Shouldn't we have then better an own directory actions instead? Yes maybe, but IMO it's not that bad to use the hooks directory. Hook extend the FAI function is some way. Defining an own action is also some sort of extension. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#391639: fai-cd creates image with wrong sources.list
Package: fai-server Version: 3.0 Tags: patch fai-cd includes the wrong sources.list to the ISO image. Therefore the image will not work in most environments. Here's the patch fai-cd @ -118,7 +118,7 @@ set +e local d -local dirs="boot $FAI media/mirror etc/fai/apt/sources.list" +local dirs="boot $FAI media/mirror etc/apt/sources.list" for d in $dirs; do umount $NFSROOT/$d 2>/dev/null done @@ -152,7 +152,7 @@ echo "deb file:/media/mirror $i $comp" >> $tmp1 done -mount --bind $tmp1 $NFSROOT/etc/fai/apt/sources.list +mount --bind $tmp1 $NFSROOT/etc/apt/sources.list -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#390367: (no subject)
I will add the new flag -u,--hostname for setting "the hostname" (only two variables) during dirinstall. But rembember that during dirinst we can set the variables HOST and HOSTNAME but FAI will not call the command hostname to change the host. If a script or command uses the command hostname to determine the hostname, if will not get the value set by this new flag. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#373596: (no subject)
This directory is normally created by make-fai-nfsroot. I will add a test into fai-chboot for testing if this directory exists. IMO fai-chboot should not created this dir. This is the task of make-fai-nfsroot. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#386192: (no subject)
The newer version of the installer only write to /etc/default/locale. /etc/environment is not used any more. We can now use the command update-locale to set these variables in /etc/default/locale. Currently LANG is set by the FAI debconf information, only LANGUAGE is missing yet. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#386192: (no subject)
The variable LANGUAGE is only set by d-i in cerain cases: Here's some code snippet from d-i: db_get debian-installer/language LANGLIST="$RET" # We set LANGUAGE only if the languagelist is a list of # languages with alternatives. Otherwise, setting it is useless if echo "$LANGLIST" | grep -q ":"; then echo "LANGUAGE=\"$LANGLIST\"" >> $DESTFILE fi atm FAI will only set LANG but not LANGUAGE. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#363701: (no subject)
A new variable was added (in FAI 3~beta60 I think). You can set FAI_ALLOW_UNSIGNED=1 in class/*.var to allow the installation of packages from unsigned repositories during the installation. When calling make-fai-nfsroot packages from unsigned repositories are allowed by default. AFAIK there are still some problems with signed repositories (e.g. when keys changing), so I will keep this behaviour until signed repositories are well supported. -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#389905: fai-chboot list wrong information
Package: fai-server Version: 3.0 Severity: minor fai-chboot -l shows wrong information for a host, if both config files like 865F09E0 and 865F09E0.disable are available. kueppers[]# fai-chboot -l [Other] 865F09E7 memtest86+ [DEFAULT] default localboot 0 demohost 865F09E0.disable install vmlinuz-install demohost 865F09E0.disable install vmlinuz-install kueppers[/srv/fai/pxelinux.cfg]# ls -l 865F09E0* -rw-r--r-- 1 lange staff-j 278 Sep 28 12:21 865F09E0 -rw-r--r-- 1 roott root282 Apr 27 11:00 865F09E0.disable -- regards Thomas -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Bug#604938: host does not boot with grub_pc and /dev/sdd as boot disk
Package: fai-doc Version: 3.4.4 In the simple examples the script GRUB_PC/10-setup always install grub into hd0. If your boot disk is not hd0, the host does not boot. I have following configuration: four SATA disks, three times 1TB, one 80GB disk. This 80GB should be my boot disk. The three disks will be part of a RAID0. During installation the disks will be named: 80 976762584 sda 8 16 976762584 sdb 8 32 976762584 sdc 8 48 78150744 sdd setup-storage creates all partitions softraid and file systems very well. (Great work Michael!). This is my disk_var.sh SWAPLIST="/dev/sdd5" BOOT_DEVICE="/dev/sdd" ROOT_PARTITION=UUID=55e150ff-0e7f-427b-ab82-0354e7515d8a BOOT_PARTITION=/dev/sdd1 Later GRUB_PC/10-setup is called without any errors = shell: GRUB_PC/10-setup = Generating grub.cfg ... cat: /boot/grub/video.lst: No such file or directory Found linux image: /boot/vmlinuz-2.6.32-5-686 Found initrd image: /boot/initrd.img-2.6.32-5-686 Found memtest86+ image: /boot/memtest86+.bin Found memtest86+ multiboot image: /boot/memtest86+_multiboot.bin done Installation finished. No error reported. GRUB_PC/10-setup OK. But the host does not boot. No grub prompt, just nothing. First, I thought is a buggy BIOS. But it's a bug in FAI IMO. Let's have a look: > cat /boot/grub/device.map (hd0) /dev/disk/by-id/ata-ST31000524NS_9WK1MYSS (hd1) /dev/disk/by-id/ata-ST31000524NS_9WK1MYEJ (hd2) /dev/disk/by-id/ata-ST31000524NS_9WK1MYRA (hd3) /dev/disk/by-id/ata-SAMSUNG_HD080HJ_S08EJ1GYC01865 My boot disk is not hd0, but hd3 (that means sdd as in disk_var.sh). But 10-setup just installs grub into hd0. > cat GRUB_PC/10-setup: #! /bin/bash # support for GRUB version 2 (1.98-1) error=0 ; trap "error=$((error|1))" ERR $ROOTCMD grub-mkdevicemap -n -m /boot/grub/device.map $ROOTCMD grub-mkconfig -o /boot/grub/grub.cfg $ROOTCMD grub-install --no-floppy "(hd0)" exit $error After manually calling $ROOTCMD grub-install --no-floppy "(hd3)" everything works just fine. A patch follows soon. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#604938: Acknowledgement (host does not boot with grub_pc and /dev/sdd as boot disk)
Here's the patch: Index: 10-setup === --- 10-setup(revision 875) +++ 10-setup(working copy) @@ -5,6 +5,8 @@ $ROOTCMD grub-mkdevicemap -n -m /boot/grub/device.map $ROOTCMD grub-mkconfig -o /boot/grub/grub.cfg -$ROOTCMD grub-install --no-floppy "(hd0)" +GROOT=$($ROOTCMD grub-probe -tdrive -d $BOOT_DEVICE) +$ROOTCMD grub-install --no-floppy "$GROOT" +echo "Grub installed on $BOOT_DEVICE = $GROOT" exit $error -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#604938: Acknowledgement (host does not boot with grub_pc and /dev/sdd as boot disk)
Package: fai-doc Severity: important I think this bug is at least important. Maybe it can causes data loss, if grub installs itself into hd0, no matter if this disk is used in a RAID. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#603675: device2grub not needed for grub2
The script device2grub is not needed for grub2. In FAI we have the class GRUB_PC which uses the script GRUB_PC/10-setup for setting up grub2 configuration. Also have a look at #604938 which fixes a problem in this script. FAI will use grub-probe (instead of device2grub) for grub2. IMO we can close this bug. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#604938: Acknowledgement (host does not boot with grub_pc and /dev/sdd as boot disk)
> - $BOOT_DEVICE may consist of several device names if RAID is used. Hence we > need to loop over each of them and do what you just proposed. That's a good reason to use your patch. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#603545: (no subject)
I contacted the tcsh-bug mailing list. Here's a reply from Jean-Luc Leger: > ok I have been tracking the problem to the function rt_mbtowc in tc.str.c > In tcsh 6.17.01, a call to mbtowc() in this function has been replaced by a > call to > mbrtowc(). > On my system, those functions behave differently when called with an empty > input string : > - mbtowc() resets the output string to \0 > - mbrtowc() leaves the output string as it is. > So, the patch should probably reset the output string to \0 before calling > mbrtowc() > or test the return value (= 0) and then reset the output string. I hope to get a patch from him, that I can test. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#603545: patch available
Thanks to Jean-Luc Leger (re...@dspnet.fr.eu.org), which created this patch. I just tested it and it fixes the bugs. tcsh-6.17.02-multibyte.patch Description: Binary data @lucas: will you prepare a new release? -- regards Thomas
Bug#736422: Is crc32 needed
I'm wondering if the kernel driver crc32 is really needed? But first, please check if dracut 034-2 (from testing, and it also works in stable) already fixes your problem. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#702019: this fix works
This bug only appears if the disk size is excatly 2500 bytes. I've tested the patch. This fix works. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#676883: partial solved
Using FAI 4.1~beta4 and dracut 034-2 I have some success. When debugging the initrd by adding rd.break=pre-pivot, I get into the initrd shell and see the the nfsroot was mounted to /sysroot via nfs v4. Also all files in /sysroot belong to user and group root. That means, that mounting via nfs v4 works, if it is properly set up on the install server. You may want to add rd.nfs.domain=localdomain to the kernel command line. After that I can say "exit", which brings me to the switch_root part. /sysroot is now mounted via aufs. But if I like to access any file in /sysroot (for e.g. passwd via cat) I always get this kind of error: switch_root:/# cat /sysroot/etc/passwd cat: /sysroot/etc/passwd: Not a directory I seems that aufs can't handle nfs v4 mounts properly. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#676883: solved!
Using the kernel from linux-image-3.12-0.bpo.1-amd64 from wheezy-backports solves this problem. Now the nfsroot can be mounted via NFS v4 without any problems. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#736470: iputils-ping: ping -b reports incorrect IP addresses when receiving multiple replies
Package: iputils-ping
Version: 3:20121221-4
Severity: important
Broadcast ping is broken.
Example where 4 hosts answer, but they are all incorrectly reported
as 172.16.1.17:
$ ping -b 172.16.0.0
WARNING: pinging broadcast address
PING 172.16.0.0 (172.16.0.0) 56(84) bytes of data.
64 bytes from 172.16.1.17: icmp_seq=1 ttl=64 time=0.597 ms
64 bytes from 172.16.1.17: icmp_seq=1 ttl=64 time=0.609 ms (DUP!)
64 bytes from 172.16.1.17: icmp_seq=1 ttl=64 time=0.610 ms (DUP!)
64 bytes from 172.16.1.17: icmp_seq=1 ttl=64 time=0.851 ms (DUP!)
64 bytes from 172.16.1.17: icmp_seq=2 ttl=64 time=0.317 ms
64 bytes from 172.16.1.17: icmp_seq=2 ttl=64 time=0.328 ms (DUP!)
64 bytes from 172.16.1.17: icmp_seq=2 ttl=64 time=0.329 ms (DUP!)
64 bytes from 172.16.1.17: icmp_seq=2 ttl=64 time=0.330 ms (DUP!)
strace shows that IP is received correctly to 'ping', but stdout write uses
wrong IP:
# strace ping -b 172.16.0.0
<...>
recvmsg(3, {msg_name(16)={sa_family=AF_INET, sin_port=htons(0),
sin_addr=inet_addr("172.16.1.132")},
msg_iov(1)=[{"E\0\0T\202\f\0\0@\1\235\365\254\20\1\224\254\20\1\3\0\0\221~\24\16\0\0015\241\341R"...,
192}], msg_controllen=32, {cmsg_len=32, cmsg_level=SOL_SOCKET, cmsg_type=0x1d
/* SCM_??? */, ...}, msg_flags=0}, 0) = 84
write(1, "64 bytes from 172.16.1.104: icmp"..., 6764 bytes from 172.16.1.104:
icmp_seq=1 ttl=64 time=0.366 ms (DUP!)
) = 67
recvmsg(3, {msg_name(16)={sa_family=AF_INET, sin_port=htons(0),
sin_addr=inet_addr("172.16.1.17")},
msg_iov(1)=[{"E\0\0T\t\10\0\0@\1\27m\254\20\1\21\214\20\1\3\0\0\221~\24\16\0\0015\241\341R"...,
192}], msg_controllen=32, {cmsg_len=32, cmsg_level=SOL_SOCKET, cmsg_type=0x1d
/* SCM_??? */, ...}, msg_flags=0}, 0) = 84
write(1, "64 bytes from 172.16.1.104: icmp"..., 6764 bytes from 172.16.1.104:
icmp_seq=1 ttl=64 time=0.488 ms (DUP!)
-- System Information:
Debian Release: jessie/sid
APT prefers testing-updates
APT policy: (500, 'testing-updates'), (500, 'testing')
Architecture: amd64 (x86_64)
Kernel: Linux 3.12-1-amd64 (SMP w/4 CPU cores)
Locale: LANG=C, LC_CTYPE=C (charmap=ANSI_X3.4-1968)
Shell: /bin/sh linked to /bin/dash
Versions of packages iputils-ping depends on:
ii libc62.17-97
ii libcap2 1:2.22-1.2
ii libgnutls-openssl27 2.12.23-8
Versions of packages iputils-ping recommends:
ii libcap2-bin 1:2.22-1.2
iputils-ping suggests no packages.
-- no debconf information
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Bug#620334: works in wheezy
My grub-probe on wheezy gives this information: grub-probe -tdrive /media/stick (/dev/sdg,msdos1) This is the correct output I expect. It seems that this a squeeze only problem and since squeeze is oldstable I like to close this bug. -- regards Thomas -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
