S print >> $TEMPFILE
echo >> $TEMPFILE
parted -s $DEVICE unit chs print >> $TEMPFILE
cat $TEMPFILE
rm $TEMPFILE
I don't really care what the partitioning is (if any at all) or whether it was
created via some setup-storage or not. I just want sample data.
thanks
thomas
t responding, and then ok, and then not responding,
I had this in the past and often the FAI client hang completely after
the NFS not responding messages. I even could not ping the
machine. But I haven't seen this the last couple of months.
--
regards Thomas
er version?
Since FAI uses the parted provided by the NFS-root it's probably safe to
ignore this oddity.
@Thomas L: What do you think about changing 'parted' into 'parted >=2.0' in
the package dependencies for fai-setup-storage and ignore parted 1.x? Even
Ubuntu 10.04 was shipping with 2.2.
bye
thomas
>>>>> On Thu, 12 Jun 2014 10:27:31 +0200, Thomas Neumann
>>>>> said:
> @Thomas L: What do you think about changing 'parted' into 'parted >=2.0'
in
> the package dependencies for fai-setup-storage and ignore parted 1.x?
Even
>
> The installing client machine has the ability to map the ip address dhcp
> gave it back to a name and the fai server can too. I'm not entirely sure
> why I see (none) there.
Did you have
use-host-decl-names on;
in your dhcpd.conf?
--
regards Thomas
his topic.
You can get an account from me if you like.
--
regards Thomas
ion without lvm, but just a
plain simple disk layout. That way, you can make sure if lvm or
something different causes the problem.
BTW, we can also switch to #fai on irc for faster help.
--
regards Thomas
> $TEMPFILE
echo "Output file is $TEMPFILE"
I'm mainly interested in lvm2
- up to 2.02.89
- 2.02.90 .. 2.02.95
- 2.02.95 .. 2.02.102
- 2.02.102+
bye
thomas
ground (FAI- and perl-wise) and what you are intending to do.
bye
thomas
On Monday 26 May 2014 11:00:29 Andreas Heinlein wrote:
> Use liblinux-lvm-perl 0.17-1 with a patch from Roland Dieterich applied
> (https://rt.cpan.org/Public/Bug/Display.html?id=94991):
I'm reasonably sure that this is the correct path. In my branch I'm now able
to feed Linux::LVM custom input.
lding the FAI packages?
In fai-doc there's the script mk-basefile which shows how to buid base
files for different distributions. I'm using debootstrap for Debian
based distributions.
--
regards Thomas
great. There's also a testing PPA
which includes FAI 4.1.1. Even testing this version would help us.
https://launchpad.net/~fai/+archive/testing/
Please report any success or problems to this list.
--
regards Thomas
fsroot and create a new pxelinux.cfg for your
install client.
--
regards Thomas
guate the $NFSROOT/etc/apt/sources.list
> properly?
In FAI 4.x (not sure if this was done before) task repository uses fcopy
for copying sources.list from the config space to the target system.
--
regards Thomas
an now say:
ainsl -s /etc/fstab 'myhomeser...
--
regards Thomas
itrd problem, or more likely a nfsroot generation
problem ?
You can join the IRC channel #fai and get quick help. Later I will we
there for more help.
I suggest to use the fai-project.org package repository, so you will
use FAI 4.2 with wheezy. That's the version I recommend for wheezy.
--
regards Thomas
so that ainsl can use this.
--
regards Thomas
27;s impossible that this could
have ever worked. As soon as you disable a volume group all related volumes
are not available any longer. Either wipe-then-disable or just-disable. I
decided for the latter.
Patch is attached, it was built against the 4.2 package. I haven't checked
what happ
On Wednesday 18 June 2014 16:59:51 Thomas Neumann wrote:
> There were 2 bugs which prevented this.
There were 3 bugs which prevented this.
c) lvremove -f $vg/$lv
Uh yes? What happened to poor '/dev/' ?
(wipefs -a $vg/$lv wouldn't have worked either, but since I removed th
>>>>> On Wed, 18 Jun 2014 16:59:51 +0200, Thomas Neumann
>>>>> said:
> +sub _make_lv_devicename {
> + # LVM 2.02.88 or below: LV Name = /dev//
> + # LVM 2.02.89 or higher: LV Name = , LV Path = /dev//
In Debian wheezy we already have LVM 2
onfig
space which also provides the GNOME desktop. They also have a mailing
list. Maybe they have already solved your issue
https://wiki.debian.org/DebianLAN
--
regards Thomas
> We (Kerim and I) are currently working on getting liblinux-lvm-perl
> patched upstream. If this does not work, we will prepare a new Debian
> version of this library.
I strongly recommend to drop Linux::LVM completely, because the required
functionality is very easily reimplemented:
e.g.:
https
On Thursday 19 June 2014 13:14:47 Andreas Heinlein wrote:
> Before your patches, we always reached the wipefs bug (missing '/dev' in
> device name). Now, on the second run, we get:
Strange. I took your provided disk_config, configured a test and got this
result:
Starting setup-storage 1.4
Using
t-get/aptitude call.
I suggest you use -v with fai-mirror and check if there's a certain
part, where the download is slow.
--
regards Thomas
move the vg1-home line and the preserve-lazy entry from my
> disk_config). These show the same error.
Then it's something with your Linux::LVM and the way it is patched. I'm
preparing a patch to replace Linux::LVM with an internal parser.
bye
thomas
needed packages?
I do not know a tools for that. But using apt-cache (depends and
rdepends) and apt-rdepends should help creating a small programm for
that task. Ask on #debian-devel if someone already did that.
I would remove the whole package list, and add every package you like to have
on this machine manually.
--
regards Thomas
ave a look at
/usr/share/doc/fai-doc/examples/simple/basefiles/mk-basefile
--
regards Thomas
Hello Andreas
Please try this patch against an unmodified 'fai-setup-storage_4.2_all.deb' and
install libcapture-tiny-perl. You can also remove liblinux-lvm-perl if you
want.
This patch is a bit larger then actually necessary since I modified the debug
output a bit and added 2 helper functions
>>>>> On Fri, 20 Jun 2014 14:29:22 +0200, Thomas Neumann
>>>>> said:
> Please try this patch against an unmodified
'fai-setup-storage_4.2_all.deb' and
> install libcapture-tiny-perl. You can also remove liblinux-lvm-perl if
you
ou should use the grub.cfg from the 4.2 packages.
--
regards Thomas
>>>>> On Sat, 21 Jun 2014 12:50:28 +0200, Thomas Lange
>>>>> said:
> Yes, in FAI 4.2 we things in fai-cd. I guess you are using the old
s/we things/we changed things/
--
regards Thomas
building two CDs, one with i386 "nfsroot", kernel and mirror and another
> one with the same for amd64.
That's a good way to do it IMO.
--
regards Thomas
t does not have this config line too.
Have a look at
https://raw.githubusercontent.com/faiproject/fai/master/conf/grub.cfg
It includes the definition of FAI_CONFIG_SRC.
--
regards Thomas
. But these do not get displayed when the script is run via
> fai softupdate.
Mmm, good question. Since fai redirects output to shell.log for
e.g. you could 'cat shell.log' after fai softupdate has run.
I'm not sure it it works when you redirect the output of your echo to
/dev/console.
--
regards Thomas
>>>>> On Wed, 02 Jul 2014 11:42:30 -0500, "John G. Heim"
>>>>> said:
> Is it redundant to do a apt-get dist-upgrade as part of a fai
> softupdate?
fai softupdates calls task_updatebase which does
aptitude full-upgrade which is the same as dist-upgrade.
--
regards Thomas
n -B. I've fixed
this now, and hope this helps. Thanks for the report.
--
regards Thomas
>>>>> On Tue, 01 Jul 2014 12:07:06 -0300, Norman Schmidt Jr
>>>>> said:
> Something I wasnt able to find out in FAI docs was how to ignore
> existing partitions when setup-storage runs.
Try setting flag_initial=1 in class/*.var
I'm not sure if this helps.
--
regards Thomas
On Thursday 03 July 2014 16:59:40 Thomas Lange wrote:
>> Something I wasnt able to find out in FAI docs was how to ignore
>> existing partitions when setup-storage runs.
> Try setting flag_initial=1 in class/*.var
> I'm not sure if this helps.
[haven't tested it out]
paring a suitable
"base.tgz" (it's just the tarred debootstrap result). [Btw. using suitable a
base image makes it possible to install RedHat and SuSE via FAI.]
bye
thomas
eed to completely rebuild the nfsroot.
bye
thomas
>>>>> On Tue, 08 Jul 2014 23:13:57 +0900, Jeong Han Lee said:
>Does anybody have some troubles with console-tools and console-common
> in NFSROOT?
In FAI 4.2 we use those packages in /etc/fai/NFSROOT:
console-common kbd
--
regards Thomas
hat log, you may understand what my problems are.
YEs, you can do that.
> I am not sure a pdf
> file can fit in this mailing list.
Please, no PDF for a log file.
--
regards Thomas
>>>>> On Tue, 8 Jul 2014 16:41:14 +0200, Holger Parplies
>>>>> said:
> Hi,
> Thomas Neumann wrote on 2014-07-08 09:36:55 +0200 [Re: packages avail
during install]:
>> On Monday 07 July 2014 17:42:58 Peter Keller wrote:
>> There
lso, the fai-monitor protocol from the client is too weak and the
> fai-monitor-gui's visualization gets garbled easily.
What do you mean by "garbled". It works very well for me.
--
regards Thomas
FAI 4.2.1 is now released. It includes several LVM related fixes in
setup-storage.
Please keep in mind that you should use liblinux-lvm-perl version
0.17-2 or newer to get all fixes. This version is included in the
package repository at fai-project.org.
--
regards Thomas
On Wednesday 09 July 2014 16:28:37 Thomas Lange wrote:
> FAI 4.2.1 is now released. It includes several LVM related fixes in
> setup-storage.
I just had a look at the patches that have been applied and noticed:
https://github.com/faiproject/fai/commit/aa5d11e0d1a93bc986e2d0c7fbd086e794
>>>>> On Thu, 10 Jul 2014 09:41:23 -0500, Peter Keller
>>>>> said:
> On Tue, Jul 08, 2014 at 05:22:36PM +0200, Thomas Lange wrote:
>> > the client machine after it finishes installing
>> Yes and no. You can use the fai-monitor for
lanation?
The message can be triggered by a physical disk, MD or LVM configuration:
https://github.com/faiproject/fai/blob/master/lib/setup-storage/Parser.pm
lines 1134+1135, 1151+1152, 1164+1165
bye
thomas
On Monday 14 July 2014 19:09:26 Holger Parplies wrote:
> If the rationale is to *force* people to edit the setting (to prevent
> putting too much load on the default server(s)),
If that is the only reason then why not use http://cdn.debian.net/ instead?
bye
thomas
ble debugging the initrd by adding rd.debug=1 to the kernel
command line.
--
regards Thomas
ipxe.org. It's the the fork of gpxe and has really nice features. For
e.g. you can boot the kernel and initrd via http. Maybe you can use
the chainloading feature of ipxe, even with your broken PXE
implementation.
--
regards Thomas
available inside the nfsroot. This is fixed in dracut 038-2 which
which is right now available in the fai-project.org package repository.
--
regards Thomas
estion: can I use the "http://fai-project.org/download";
also for Ubuntu 14.04/trusty? (I remember having heard something in former
e-mails)
Yes, give it a try.
Someone posted on this list that the version 4.1.1 from
https://launchpad.net/~fai/+archive/ubuntu/testing works, so I guess
FAI 4.2.1 from the fai-project repository should also work.
--
regards Thomas
re redirected to is
currently broken.
--
regards Thomas
repository, after you've added this line to sources.list:
deb http://fai-project.org/download wheezy koeln
--
regards Thomas
o do that in the new dracut world order.
man dracut.conf, then look at
add_drivers+=
--
regards Thomas
nts after installation
b) configure dhcp server and use dhcp to assign a static MAC<->IP mapping
c) install via fai-cd (actual cd or flashdrive)
'b)' can be used to write a static network configuration during installation
I think that's about it.
bye
thomas
4GiBext4 rw,nodev
VolGroup00-LVvar /var4GiBext4 rw,nodev
(I also replaced all raw numbers with appropriate units and configured the
correct minimum size for the LVM PV. Additional tip: if this is a VM configure
/tmp and optionally /var/tmp as tmpfs devices to improve performance.)
tschüß
thomas
d yield the expected
values.
a) can someone confirm his setup_storage generates the same strange values?
b) can someone explain what kind of bash trickery is going on here?
bye
thomas
better in this case. the shorter ${PARAMETER:=WORD}
> would assign the value of WORD to PARAMETER in the unset/empty case. in
> my head ":-" is for returning defaults, ":=" is for assigning them. but
> please correct me if i'm off-base, as i'v
On Wednesday 06 August 2014 09:26:08 andrew bezella wrote:
> On Wed, 2014-08-06 at 13:11 +0200, Thomas Neumann wrote:
> > P.S.: @Andrew: Do you know where 'grok' is coming from?
>
> it's from mars :)
>
> brought to earth by valentine michael smith in hein
i-ish way to accomplish any required changes to these
variables.
I have prepared a patchset for Thomas & Kerim which moves the actual file-
content generator logic for disk_var.sh from setup-storage to Fstab.pm (to
get rid of generating the output for debug and the actual file content
ist.
Die Konfiguration sieht gut aus und ich habe sie eben erfolgreich mit einem
setup-storage 1.5 (aus fai 4.0.8) durchlaufen lassen.
Your configuration looks good and I have successfully applied it using setup-
storage 1.5 (from fai 4.0.8).
bye
thomas
nachgebaut und setup-storage
1.5 läuft erfolgreich durch.
tschüß
thomas
P.S.: '1-100%' bedeutet '1MiB bis 100%'. Ja das ist nicht intuitiv.
switch back to MSDOS partitioning?
b) From the top of my head:
"disk_config sda disklabel:msdos"
bye
thomas
orow. I will report problems if any.
I am pretty sure this is actually a SecureBoot / UEFI problem in disguise.
bye
thomas
in an upcoming release.
Error: Could not detect file system.
bye
thomas
complains)
bye
thomas
On Wednesday 20 August 2014 22:35:46 Andrew Ruthven wrote:
> On Wed, 2014-08-20 at 11:19 +0200, Thomas Neumann wrote:
> Nope, this is from GPT. For example, in parted, use the "name" command
> to set a name on the partition. Below is an example (assigned manually
> after a
ed packages).
bye
thomas
ing into 40-install-grub:
set -xv
Please look at the output of shell.log, which has more information
than error.log
--
regards Thomas
for Debian
Jessie (8.0). This is the sources.list entry for it.
deb http://fai-project.org/download jessie koeln
--
many greetings from DebConf in Portland, OR
Thomas
ow one
you can speed up installation for all OS variants.)
bye
thomas
r this constellation?
Yes. Please add this info to the wiki too.
--
regards Thomas
>>>>> On Tue, 29 Jul 2014 16:20:22 -0500, Peter Keller
>>>>> said:
> /etc/dracut.conf.d/whatever.conf file. Then, how do I run dracut
> such that it reproduces the initrd.img.blah in the /boot chroot
> which
chroot /srv/fai/nfsroot dpkg-reconfigure dracut
--
regards Thomas
ttp://fai-project.org/download jessie koeln
to your sources.list files.
--
regards Thomas
/tmp 5Gext4 rw,noatime,nosuid,nodev createopts="-L tmp -m 0"
tuneopts="-c 0 -i 0"
primary /usr 6Gext4 rw,noatime,nodev createopts="-L usr"
primary /files-scratch10%- ext4 rw,noatime,nosuid,nodev createopts="-L
storage -m 1" tuneopts="-c 0 -i 0"
With gpt you do not have logical partitions.
--
regards Thomas
can do following things:
ssh-keyscan installserver > /srv/fai/nfsroot/root/.ssh/known_hosts
ssh-keyscan IP-of-installserver >> /srv/fai/nfsroot/root/.ssh/known_hosts
Then the host key is already know inside the nfsroot.
cp ~fai/.ssh/id* /srv/fai/nfsroot/root/.ssh
This is for not asking the pw.
--
regards Thomas
.debian.org/cgi-bin/bugreport.cgi?bug=764313#10, where I
explain why this is happening.
--
regards Thomas
t.org/download jessie koeln
--
regards Thomas
ot_ does not exists
Please check if you have rd.auto or rd.auto=1 on the kernel command
line. This is needed so that dracut activates LVM.
--
regards Thomas
f packages are missing, you should ad the to
/srv/fai/config/packages_config/CENTOS.
Maybe add 'base' to the section yumgroup.
--
regards Thomas
get and call lsinitrd to see if
the lvm commands are in the initrd.
If not, you can tell dracut which "modules" to add into the
initrd. See dracut.conf for more infos and dracut.cmdline for
debugging help.
--
regards Thomas
ell.log: WARNING: Failed to connect to lvmetad: No such file or
directory. Falling back to internal scanning.
You should read the whole shell.log
--
regards Thomas
this once you have installed the FAI packages.
When builing the nfsroot /etc/fai/apt./sources.list
^^^
is used. There you should include the FAI repository (for getting the
newer FAI packages).
--
regards Thomas
for wheezy.
Since the breakage is in Ubuntu, I don't know how to fix this. I'm also
wondering why Ubuntu 14.04 only has console-setup 1.70. Even Debian
wheezy (released in may 2013) has version 1.88.
Maybe you can fix this by setting the initramfs-tools package on hold.
--
regards Thomas
remove
$target/etc/apt/apt.conf.d/02proxy at the end of the installation.
P.S: I'm always very interested in user experiences with FAI. Can you
please fill out the FAI questionnaire
http://fai-project.org/questionnaire
and send it back to me when you finished a project with fai.
--
regards Thomas
really a bad change. I fixed this in
adb8ce8ec48a5564bfa0744ac672. I like to have opinions if this first
change is a important bug. Currently we do not have a bug report for
this. Do we need to fix it for jessie? I guess the next official FAI
version will no make it into jessie, because if has more changes than
bug fixes only.
--
regards Thomas
ages. It builds a complete new
package repository structur. Is there a tool that can do it better?
--
regards Thomas
introduced in FAI 4.2.2.
The 4.3.1 packages are available for jessie. I will try to prepare a
similar package for wheezy.
--
regards Thomas
I've created FAI version 4.3.1+wheezy1, which is the same as 4.3.1
except that not jessie but wheezy is the default distribution.
It's available now if you use this line in your sources.lists files:
deb http://fai-project.org/download wheezy koeln
--
regards Thomas
Uhm. Don't do that. The canonical solution is to mount sda1 directly as a
normal filesystem and manually sync sda1 to sdb1 after a successfull boot.
Everything else is creating a dependency problem between initrd, md and
whatever filesystem you're using. It could be wrangled into a working by
usin
to get a software-raid /boot to work.
But it has 2 distinct advantages:
- if something accidentally deletes / corrupts a file in /boot there's
still a reasonable chance for recovery _during system boot_
- it works with all combinations of fat/gpt/md/lvm/crypt/btrfs
bye
thomas
The new FAI CD image is available. It's made with FAI 4.3.1+wheezy1
and can install a simple demo host, a host containing XFCE
enviroment. It also includes the rescue boot option and installing a
FAI server. The ISO image is available from
http://fai-project.org/fai-cd
--
regards Thomas
>>>>> On Mon, 1 Dec 2014 10:18:23 +, Fabian Hartmann
>>>>> said:
> I’m trying to setup a Ubuntu Linux with fai-cd.
Which FAI and which Ubuntu version are you using?
--
regards Thomas
ce to this? I thought about a little ncurses menu where
the user can enter the IP, netmask and gateway.
Yes. Have a look at
http://fai-project.org/download/misc/30menu.source
--
regards Thomas
which is used during installation, which
may differ from the version you have installed on your install server.
--
regards Thomas
e system
After rethinking this issue I think it can happen, if you do not have
defined a separate /tmp partition in disk_config. Is this the case in
your environment?
--
regards Thomas
mkdir -p $target/tmp
mount -t tmpfs tmpfs $target/tmp
--
regards Thomas
How can I configure the live-system to provide me with shell-access in
another terminal. I remember to have had this while using fai with Debian.
Normally pressing alt-F2 or alt-F3 gives you a console, if you have
defined the FAI_FLAGS=createvt
--
regards Thomas
701 - 800 von 2000 matches
Mail list logo