27;gpt-bios'
should not be a valid keyword: disklabel:(msdos|gpt)
Additional question 2)
Is there a way to make setup-storage create the gpt bios partition as
partition 1
instead of 3? Rationale: Nothing technical. Just a cosmetical issue:
It looks better
to me.
bye
thomas
>>>>> On Mon, 19 Mar 2012 23:32:51 +0100, Thomas Neumann
>>>>> said:
> Does the setup-storage tool try to create properly aligned GPT
> partitions? Reason for asking is a somewhat discouraging information on
> wikipedia:
There's the opt
<- sincerely apologizing for the broken linewraps in the previous posts
I think it's finally time to scrap this wannabe 'mail client'.
> the reasoning from /usr/share/fai/setup-storage/Sizes.pm:
>
> # on gpt-bios we'll need an additional partition to store what
> # doesn't fit in the MBR; this partition must be at the beginning,
> # but it should be created at the very end such as not to invalidate
> # indices of other partitions
ding the basefiles via http. This is already included
in the beta version of FAI 4.0.
> The extrbase hook? Or is there a better place?
Very good idea.
--
regards Thomas
be used for creating the partitial
mirror for the FAI CD. But there's already a wishlist bug to extend
this script for using this mirror for the network installation and for
creating the nfsroot.
--
regards Thomas
>> The extrbase hook? Or is there a better place?
> Very good idea.
In a different country and a different FAI installation I used this
"hooks/extrbase.DEFAULT" in place of the 'normal' one. The hook assumes
an environment variable ${BASEFILE} or
${DEBOOTSTRAP}/${DEBOOTSTRAP_OPTS} exists and d
partition
(e.g. /dev/sda3) but the first 'physical' partition (e.g. sector 63-...)
but gfdisk and parted tell me otherwise.
So - what's the proper way to achieve it? It's okay if I have to switch
the nfsroot to Debian and go experimental or something like that. I'm
quite comfortable with handling differing nfsroot and client OS anyway.
bye
thomas
Hello everyone
a simple gpt-enabled disk_config
disk_config disk1 disklabel:gpt
primary-1 - -
primaryswap 8Gswapsw
primary/4G- xfs rw
results in
SWAPLIST="/dev/sda2"
BOOT_DEVICE="/dev/sda"
ROOT_PARTITION="/dev/sda3"
BOOT_PARTI
Ich wünsche allen Lesern und Lurkern frohe OsterFAIertage.
SCNR.
-u localhost nfs 3 >/dev/null 2>&1 ||
RPCMOUNTDOPTS="$RPCMOUNTDOPTS --no-nfs-version 3"
afterwards test with 'rpcinfo -u localhost nfs 3'
expected result:
program 13 version 3 ready and waiting
bye
thomas
?
There's more than one solution. Did you read this?
http://wiki.fai-project.org/wiki/Multiple_Network_Interfaces_Howto
And this?
http://www.mail-archive.com/linux-fai@uni-koeln.de/msg04452.html
--
regards Thomas
e setup-storage is now included in a separate package, it should
be easy to update just this package.
--
regards Thomas
, if you have defined LOGSERVER and LOGUSER the install client
calls fai-chboot -d and during the next reboot, the client get the
default pxelinux configuration which says localboot 0.
--
regards Thomas
ation?
The option -P will be removed in FAI 4.0.
--
regards Thomas
em to guard the enclosed
code against running during a softupdate.
a) What's the difference between these two? Is there any?
b) What is the proper/best/most obvious/recommended way to let code run
only during the initial installation?
c) Does install/dirinstall make a difference?
bye
thomas
oment I try to force-feed a working kernel boot config into
grub.conf manually, but I really like to do it properly. I'm not confident
that my manual intervention will survive the next kernel-update.
bye
thomas
>>>>> On Tue, 22 May 2012 12:56:56 +0200, "Thomas Neumann"
>>>>> said:
> Hello
> In the original prepareapt hook, there are (were?) two different
if-clauses:
> if [ -f /var/run/fai/FAI_INSTALLATION_IN_PROGRESS ]; then
> if
her type of "initial" inside FAI. It's only disk
setup-storage related.
--
regards Thomas
>>>>> On Tue, 22 May 2012 15:54:08 +0200, "Thomas Neumann"
>>>>> said:
> Maybe someone has already done this. I'm currently a bit at a loss on how
> to manually configure a RedHat (or more specifically OracleLinux 6) grub
> boot
> Then if I try to do and install pxe loads then a lot of text flies by and
> then see
> Begin: Loading essential drivers ... done
> Begin: Running /scripts/init-premount ... done
> Begin: Mount root file system ... /init: .: line 249: can't open
> 'scripts/live'
Your nfsroot may not have the
l pxe/samba/nfs/dhcp server is still running lenny.
> Q1. Is the .deb packages above compatible with lenny?
No.
> Q3. Are there any documentation or howto on this subject?
There's the wiki and the FAI guide and a lot of other documentation at
http://fai-project.org/otherdocs/
--
regards Thomas
>>>>> On Thu, 24 May 2012 16:53:25 +0200, "Thomas Neumann"
>>>>> said:
> - nfs/aufs related quirks -
> c) aufs does not work. Tells me something is not right (/etc/fstab,
> /sysconfig/etc/fstab is missing) and gives up with a a k
root=11.22.9.158:/srv/fai/nfsroot:vers=3 rd.shell
FAI_FLAGS=verbose,sshd,createvt FAI_CONFIG_SRC=nfs://11.22.9.158/srv/fai/config
FAI_ACTION=install rd.md=0 rd.dm=0 rd.lvm=0 rd.luks=0 vga=794
There's no need to disable nfsv4 on the install server.
--
regards Thomas
tion and FAI is using it for several
> years. But if you do like to do it without nfs+aufs read
> http://fai-project.org/doc/man/fai-nfsroot2image.html
> It's working like a charm.
Thank you for both hints. I will try it as soon as I have a bit of time!
Bugreport: Surely.
bye
thomas
t;ERROR: Set $FAI_CONFIG_SRC
> or use option -s"
> In fai version 3.8 everything worked fine
> configuration files of version 3.8 not work in version 4?
You should read the NEWS file. This describes important things that
have changed.
--
regards Thomas
something went wrong during your last fai-make-nfsroot
run. Check which directories are mounted (like /srv/fai/nfsroot/proc,
/sys, /dev and) and umount those before rerunning
fai-make-nfsroot.
--
regards Thomas
if nothing is mounted. The call fai-make-nfsroot -v and put
the log file onto paste.debian.net. I will have a look at it then.
P.S.: Please do not cite the whole mail if not necessary.
--
regards Thomas
t: failed to run command `aptitude': No such file or directory
I think you've hit this bug.
https://bugs.launchpad.net/ubuntu/+source/fai/+bug/714570
--
regards Thomas
Thomas Lange wrote:
> I tested FAI 4.0 and dracut inside a wheezy nfsroot on a squeeze
> install server. This works fine. I can reproduce your errors on a
> wheezy install server. I will have a look into it. Maybe you can write
> a simple bug report for this.
I filed Bug#676882 (L
ocess:
> FAI_CONFIG_SRC is set to nfs://server/srv/fai/config
> I get the following error:
> mount.nfs: mount system call failed
Strange. Can your install client resolve the IP address of server?
Does your install server have multiple network cards?
--
regards Thomas
>>>>> On Mon, 11 Jun 2012 15:56:11 +0200, steph steph
>>>>> said:
> I saw on a forum that someone tried and for him it seems to work but when
i apply the configuration to
> my FAI deployment doesn't work.
Any error messages? What exactly does not work?
--
regards Thomas
First we have to make sure that this causes no
errors.
--
regards Thomas
n 3"
The option --no-nfs-version 3 is only used for the mount daemon, it's
not used for disabling nfs v3 by default. And this is only set after
asking the portmap or rpcbind service if nfs v3 is available on
localhost. IMO this is perfect fine and I see no reason why it should
break FAI.
--
regards Thomas
left and right, but in the end it
was a simple file that was missing. *gnarf*
touch /srv/fai/nfsroot/.THIS_IS_THE_FAI_NFSROOT
fixed that. Since I was only tarring up '[...]/*' the hidden file didn't
make it into the tarball and fai refused to proceed when I was booting the
client.
tschuess
thomas
has not started at this time, it looks more
> to me like a kernel bug. FAI 4 has been working fine for me before this
> bug has shown up, and is still working on i386 arch with the same
> configuration (Server is using Squeeze, and uses multiple nfsroots).
Are you using dracut or live-boot inside the nfsroot?
--
regards Thomas
[This is a repost of a message I sent earlier. It contained 2 screenshot
attachments and was ignored by the ML-software due to its size.]
> I filed Bug#676882 (LVM) [...]
I now have the option to decide between a n[e|o]tbooting dracut-faiclient.
If I do not provide rd_NO_LVM via tftp, then the
> If I do provide rd_NO_LVM, then fai starts (finally), but setup-storage
> bugs out rather drastically [compare evidence b) ]
>
> http://www.fluffbunny.de/fai-dracut-lvm-1.jpg
sorry. copy&paste error. correct link is
http://www.fluffbunny.de/fai-dracut-lvm-2.jpg
> What would be the best path to follow? Mirror the whole disks and put
> all the partitioning onto the md0 device? Or create partitions on both
> disks independently and then pair these up in RAID? At what level is it
> best to introduce LVM?
For _GRUB1_ configurations my typical suggestion is:
age), DHCP, the kernel command line and
FAI itself.
Check boot.log which contains the result from a DHCP query inside FAI.
Normally FAI disables a lot of options in live-boot with the config
file /etc/fai/live.conf. Have a look at it.
--
regards Thomas
> where is the hostname set during the FAI boot?
I ~think~ it's just the DHCP server that set's the name. The dhcp-client
can be told to ignore this setting. Then you would end up with whatever is
currently configured (Debian/Ubuntu: /etc/hostname, SuSE: /etc/HOSTNAME,
RedHat: cat /etc/sysconfig/n
> Find the row and add HOSTNAME=$host:
> $append="append $initrd $bootprot $rootfs HOSTNAME=$host $opt_k $flags
> $action\n";
What does the nfsroot's `hostname` command report as the current hostname?
Is it the same as the content of HOSTNAME or is it the original/unmodified
host's name?
>> Find the row and add HOSTNAME=$host:
>> $append="append $initrd $bootprot $rootfs HOSTNAME=$host $opt_k $flags
>> $action\n";
> What does the nfsroot's `hostname` command report as the current hostname?
> Is it the same as the content of HOSTNAME or is it the original/unmodified
> host's name?
7;safe side' by explicitely next-ing via a label.
I like labels. Code-wise there's no difference but it's nice if the code says
what it is looping over. (I hope I got it right it's a command. On first
reading this code fragment I thought it was a dependency.)
tschüß
thomas, trying
/root -- system left untouched.'
If this is the correct error, then I ~just~ need to find out why
'/dev/vg_system/root' has noot been created.
tschüß
thomas
--
working FAI 4.0 installation with dracut-nfsroot and LVM
disk configuration. Simple disk partitioning without LVM works too.
Attached is the full patch for 'version 2)'.
tschüß
thomas
Commands.pm_patch
Description: Binary data
efore task partition is called. First use
fai-mount-disk -f
for mounting the local disk according to the fstab. Then check if
there are files in $target/home. If files are found, stop the
installation, otherwise call fai-mount-disk -u to umount the local
file systems and go on with the installation.
--
best regards Thomas
that point though. I just described how
to prepare the base server.]
tschuess
thomas
P.S.: I have successfully installed SuSE with older versions of FAI and it
was a bumpy road. (Most trouble stemmed from properly configuring the boot
loader.) Some changes have been made in 3.4.8, but I haven&#x
>>>>> On Tue, 26 Jun 2012 23:25:28 +0200, Oskar Ahner said:
> However, my question is: Is it possible to run the fai-server on Suse, or
is it bound to Debian ?
As Thomas already pointed out, it possible with any unix-like server
but it's much easier to set up
ueeze and Wheezy, all x86 and
amd64. Is the new version
> incompatible with any of these? As I said the server is, and will be
running Squeeze.
We dropped lenny support.
--
regards Thomas
flags are: FAI_FLAGS=verbose,sshd,createvt,debug
(Installation mode is network install with PXE boot)
Has this changed? The changelog doesn't seem to mention this.
tschüß
thomas
't work as expected.
Maybe I never had a faultless installation before? Now that I think about
it, there was this pcspeaker module load error which I simply ignored.
@Thomas:
Is it too late to request some behaviour change for 4.x?
Something along the line of:
- system will not re
Attached is a small hook which just makes sure the system is not
rebooted/halted if no error occured AND no reboot/halt flag is set. Maybe
someone finds this useful. (e.g. for debugging/inspecting an installation.)
Tested only with initial network installation. I don't know how well it
interact
>>>>> On Mon, 9 Jul 2012 18:50:54 +0200, Thomas Neumann
>>>>> said:
> @Thomas:
> Is it too late to request some behaviour change for 4.x?
YUes, it's too late. We had a long testing period which ended with the
upload of 4.0.
> Somethin
or
> http://paste.debian.net/179755
> http://imageshack.us/photo/my-images/191/faidracut.png/
Something weird with your kernel. Why doe this kernel doesa not
support nfs?
> I've also tried with live-boot
When using live-boot you must not install dracut into the nfsroot. And
you also need the lkernel parameter boot-live.
--
regards Thomas
ling rsyslog with
# "start: Unable to connect to Upstart: Failed to connect to
# socket /com/ubuntu/upstart: Connection refused"
--
regards Thomas
Here's a better workaround for it.
wheezy[~]> diff -u ~/fai-4.0.3/bin/fai-make-nfsroot
--- ~/fai-4.0.3/bin/fai-make-nfsroot 2012-06-26 16:22:27.715039242 +0200
+++ /usr/sbin/fai-make-nfsroot 2012-07-20 00:22:21.025226452 +0200
@@ -452,7 +452,7 @@
done
fi
$ROOTCMD apt-get updat
d for syslog logging
--
regards Thomas
dung"chroot: failed to run command
> 'apt-get': No such file or directory" ab. Das angelegte Log enthält auch
Das deutet immer auf einen kaputten Debian Mirror hin. Versuche es mal
mit einem offiziellen Mirror.
--
regards Thomas
rs, then fai _will_ reboot the system. You can tell fai to halt
instead, but you can't simply make it stop and do nothing. Reboot or Halt.
Solution:
a) provoke an error somewhere during execution
b) use my attached patch to force a non-reboot
bye
thomas
g "reboot" was set, or
> sit there if not).
I checked the recent 3.x versions and it wasn't changed. However at least for
me there always occured an error regarding the pcspeaker module, so if I didn't
specify 'reboot' it would wait for a confirmation ('Press Enter to continue').
bye
thomas
taining the
software, but I will not work on the documentation.
--
regards Thomas
oot does not use debconf preseeding.
We do have NFSROOT_HOOKS, but IIRC they are called only at the end of m-f-n.
Which packages are affected by this?
--
regards Thomas
manual in the
examples.
--
regards Thomas
nel command line.
Use fai-chboot -u for this.
> The problem might be solved by adding the appropriate line in fai.conf
> in the chroot.
This is deprecated and not supported any more.
--
regards Thomas
ons.
> I don't see how to use FAI with a specific iso - Wheezy Beta 1.
FAI does not use an ISO. It uses packages from a Debian repository
which are defined by the sources.list file as usually.
--
regards Thomas
hiya
Brian Kroth wrote:
> Let me know if you need any more details.
I'll have a look into this.
About the perl prototype error: Encountered the same one, but ignored it since
it seemed to have no impact.
bye
thomas
esting. Will it replace initramfs-tools/live-boot
in
> the future?
Yes, dracut is the default when calling fai-make-nfsroot in FAI
4.x. fai-cd is not yet working with dracut, but it shouldn't be that
hard to fix this.
--
regards Thomas
>>>>> On Tue, 21 Aug 2012 16:57:24 +0200, Reinhard Tartler
>>>>> said:
> Thomas, can you please give me a pointer about how dracut is
> integrated in fai-make-nfsroot?
It's just a Debian package which is installed into the nfsroot.
Adding t
less important in FAI. Currently I'm very
busy but I think in october I will have much more time to work again
on FAI.
--
regards Thomas
s attached patch, which simply logs and
ignores any 'self_cleared_[a-z.]' dependencies. This makes the resolver finish
successfully, but afterwards there's this little gem:
wipefs -a vg_system/fscache
Command had non-zero exit code
Erm. 'vg_system/fscache'??? The previous
x27;[fake-mode] $FAI::commands{$_}{cmd} . "\n";
That way you can see the list of commands and have a visual check on what
is supposed to be done. Check if anything strange is executed or the order
is messed up.
bye
thomas
Hello Brian
> With your patches, an initial install of the attached MD_RAID1 + LVM
> disk_config works, but a subsequent preserve_reinstall (without initial
> flag) fails with a non-zero return code from "Executing: mdadm --stop
> --scan". Full output is attached.
[...]
> Might have something
Thomas Neumann wrote:
> I extracted a commandlist from your pastebin. Seems like the failing
> command is the first one. Would be interesting to know which exit code
> has been signaled. Try modifying Exec.pm::execute_command() but you may
> have more luck kludging some output hand
>>>>> On Mon, 27 Aug 2012 21:27:43 +0200, Thomas Neumann
>>>>> said:
> Turn
> &FAI::execute_command($FAI::commands{$_}{cmd});
> into
> print '[fake-mode] $FAI::commands{$_}{cmd} . "\n";
> That way you can s
> Why not use setup-storage without -X which prints all commands without
> executin them?
Ah. Yes. Forgot about it since I started rewriting setup-storage so it
doesn't need root-priviledges for a dry-run. (e.g. setup-storage without
'-X' will still happily try to stop md devices and access the ra
o say it
would be pretty nice if one doesn't actually have to care about this at
all.
bye
thomas
Thomas Lange wrote:
> Why not use setup-storage without -X which prints all commands without
> executin them?
Okay, I tried running with '-X' now, but there's still a logical bug which
makes it a bit less useful for automated testing[1]. After the commands
have been execu
Hello everyone
Can you help me out and send me your disk_configs (via private mail)? The more
the better. May it be a trivial one or highly sophisticated. I'd like to build
a nice sample collection which to test setup-storage against. [It doesn't even
matter if the config is supposed to work, b
This is a thank you for all the people who have sent me their configs. It's
truly amazing what you have come up with or which outside constraints you had
to deal with and came up with a solution. =)
rable qw(dclone);' somewhere around line
696 and 717. Move one of these lines to the head of the file, right next
to 'use Parse::RecDescent;' and delete the other one. That's it already.
(Perl just didn't like having the same file included twice.)
bye
thomas
big help, but I now find
it a bit lacking. I'm currently pondering on turning this bash beast into
a nice perl program with the goal of being easily extendable with anybody's
personal disk config. If nothing else it should provide a nice training to
hone my programming and design skills. =)
bye
thomas
ly replaces
Kerberos, maybe puppet can be used to distribute the keytab (and other
configuration items).
bye
thomas
better solution?
Convince setup-storage that 'preserve_???' really means 'do not touch this
partition at all?
tschüß
thomas
P.S.: Uh. Oh. I can see where this is heading.
g /boot/initrd.img-3.2.0-3-amd64
E: No '/dev/log' or 'logger' included for syslog logging
F: Failed to install /etc/idmapd.conf
dpkg: error processing dracut (--configure):
Seems to be
t any problems. NFS is not a problem,
when doing installations with FAI. Maybe your network card is
fully stretched and therefor the server seems to be unresponsive.
--
regards Thomas
NFS
parameters from cat /proc/mounts
1.2.3.149:/srv/fai/nfsroot-squeeze64 /live/image nfs
ro,relatime,vers=3,rsize=1048576,wsize=1048576,namlen=255,hard,nolock,proto=tcp,port=65535,timeo=7,retrans=3,sec=sys,mountport=65535,addr=1.2.3.149
0 0
IMO there's no need to set the rsize parameter.
--
regards Thomas
also use them to 'skip'
the debootstrapping stage with Debian/Ubuntu.)
bye
thomas
MAC address flooding.
IMO the only really secure way is to enter some secret on every
machine, for creating a secure and encrypted communication channel to
the install server. But even then, you have to trust into the person
how enters the secret on every machine.
--
regards Thomas
systems. Tar'ing and copying 500MB of data gets old quite fast.]
bye
thomas
sysinfo
I also ran fai-setup -f shortly before this, so it's up-to-date.
Thank you for any help,
Thomas Koch, http://www.koch.ro
>>>>> On Mon, 5 Nov 2012 16:09:19 +0100, Thomas Koch said:
> No I get on the client:
> nfsroot type nfs requested but kernel/initrd does not support nfs
Never seen that. Do you use the plain Debian kernel?
My pxelinux.cffg for wheezy is like this:
default fai
The FAI web site will be down from today 5pm until saturday evening.
--
regards Thomas
I'm wondering if anyone can help with #682067. Please write a comment
to the bug report. I'm interested in confirmation of the bug, or
comments if you do not encounter this problem.
--
regards Thomas
the install server. Specifying it
here with an IP address is not recommended.
--
regards Thomas
in nfsroot.conf to
NFSROOT=IPADDRESS:/srv/fai/nfsroot, which
> works fine.
> But now i got " no supported filesystem images found at /live". It seems
i have trouble
> with the UnionFS, it is not writeable. Any ideas/tips
Same problem as before. Do not add the IP address to the variable
NFSROOT in fsnroot.conf.
--
regards Thomas
the current version of live-boot in testing is not working
properly, I've put a more recent copy of the live-boot package (from
unstable) into the fai-project.org repository.
--
regards Thomas
o long. Also disabling Spanning Tree Protocol
(STP) on your switch ports via PortFast (on Cisco) may help.
You can also switch your nfsroot from using live-boot (which uses
klibc which uses ipconfig) to dracut. I recommend this for FAI 4 with a
wheezy nfsroot.
--
regards Thomas
FAI 4, ainsl uses the variable AINSL_TARGET which is set by
default. Therefore I had to remove $target from my config script from
each ainsl call like this
ainsl -v /etc/hosts
After that, ainsl works as expected.
--
regards Thomas
A new version of FAI is now available. No big changes. Changes can be
found at
http://packages.debian.org/changelogs/pool/main/f/fai/current/changelog
--
regards Thomas
401 - 500 von 2000 matches
Mail list logo
