Hi,
We noticed we had problems with servers and workstations not restarting
services, when upgrading packages, as they should. We learnt that the small
shell script /usr/sbin/policy-rc.d was the culprit.
I tried to remove it with a $FAI_CONFIGDIR/scripts/LAST/99-no_policy-rc.d
script, but I
On 05/11/11 11:11, Michael Tautschnig wrote:
Given that 3.4.8 has been released which includes all the bugfixes up to
experimental68 and several more, it might be good for all production systems to
go for that version, although further testing will definitively required to rule
out problems such
On 05/10/11 19:20, Sylvain Milot wrote:
I then tried the experimental version - version 4.0~beta2+experimental60 at the
time
and this worked quite nicely. I then recently rebuilt my NFSROOT with the
latest experimental
version and it now complains about overflow. The behavior is the same whethe
On 03/16/11 15:31, Michael Tautschnig wrote:
Hmm, ok, --wait-clean isn't supported in lenny. Could you give experimental68
another try? I've removed that option now.
You hit it! 1k mega-points to you, sir! =)
I've tried Squeeze and Lenny installs on both a workstation, RAID+LVM
and preser
On 03/16/11 13:47, Michael Tautschnig wrote:
Oh, yes, of course, sorry. The point is that the built-in mdadm-startall was
running too late to have the LVM on top of RAID detected. Could you please give
4.0~beta2+experimental67 another try *without* your manual mdadm-startall? That
version finally
On 03/16/11 10:53, Fredrik Eriksson wrote:
I'm sorry, I must have missed rebuilding a nfsroot when testing
yesterday. Last successful RAID+LVM w/ preserve was with
4.0~beta2+experimental61, I noticed when looking at log files.
Unfortunately it did not work, neither with nor without the
Hi,
I'm sorry, I must have missed rebuilding a nfsroot when testing
yesterday. Last successful RAID+LVM w/ preserve was with
4.0~beta2+experimental61, I noticed when looking at log files.
On 03/15/11 18:34, Michael Tautschnig wrote:
I don't know whether Fredrik Eriksson will love
On 03/15/11 17:53, Michael Tautschnig wrote:
Thanks a lot for sending the logs as well (was done in private mail, just in
case anybody wonders). I hope to have addressed this in 4.0~beta2+experimental65
- I couldn't test it though as the test system is at home, but at least there
were obvious bug
Hi,
A new problem arose when trying to install a one disk system today, and
we think it boils down to this:
Trying to add CMD: lvcreate -Cannot satisfy pre-depends for pvcreate
-ff -y /dev/cciss/c0d0p2: ,pt_complete_/dev/cciss/c0d0 -- system left
untouched.
There probably shouldn't be a
On 03/14/11 18:33, Michael Tautschnig wrote:
I was trying to reproduce this situation over here for further testing. Using
your config I first created the RAID array and the LVM on top of it. After a
manual shutdown using vgchange -a n ; mdadm --stop --scan not even doing
mdadm-startall would rev
On 03/11/11 17:20, Michael Tautschnig wrote:
Thanks a lot for letting me know! Please note, however, that I fixed one or two
bugs and it might be worthwhile updating to 4.0~beta2+experimental60 (I think
that's the latest version).
We're now at 4.0~beta2+experimental61.
I ran into something
On 03/10/11 15:10, Michael Tautschnig wrote:
As of 4.0~beta2+experimental57 setup-storage is finally capable of doing that
and should properly handle arbitrary stacking of LVM/RAID devices. That said,
testing is of course much appreciated :-)
Today I've had several successful re-installs wi
On 02/08/11 13:21, Michael Tautschnig wrote:
I think this problem should be solvable by installing mdadm on the system,
because this only will trigger adding md support in the initrd. Mind adding
mdadm to a suitable package_config/* file and retry?
We actually had that in our package_config
Hi, Michael
Thanks for your reply. I am trying out a few things you mentioned and
will send you a few log files privately.
Michael Tautschnig wrote:
Just a note: it shouldn't be necessary to add the preserve_lazy:1,2 and
preserve_lazy:0,1 lines - setup-storage should definitely be able to fi
Fredrik Eriksson wrote:
I'll attach two log files. The format.log is from an install without
modifications to usr/sbin/setup-storage and format2.log is from an
install with the modifications mentioned in the diff above.
Uhm, right.. =)
--
Fredrik
formatlog.tar.gz
Description: GN
to boot the system.
vg0-home is not preserved since the RAID isn't initialized prior to
get_current_lvm.
Fredrik Eriksson wrote:
when using these modifications
--- /srv/fai/lennyroot64/live/filesystem.dir/usr/sbin/setup-storage
2010-12-23 23:21:37.0 +0100
+++ /srv/fai/squeeze
Fredrik Eriksson wrote:
So, I've updated our production ready system to 3.4.7 and installed
4.0~beta2+experimental49 on a test server. Both seem to work fine with
our current config.
The following mail concerns tests made with the 4.0~beta2+experimental49
FAI.
I'll get back on th
I'm back! =)
Michael Tautschnig wrote:
Could you give 4.0~beta2+experimental48 a try on this test system of yours? I've
already asked Mât and Nicolas to test this version as you might have just seen,
but the more testers we get, the better :-)
So, I've updated our production ready system to
Fredrik Eriksson wrote:
Michael Tautschnig wrote:
Could you give 4.0~beta2+experimental48 a try on this test system of yours? I've
already asked Mât and Nicolas to test this version as you might have just seen,
but the more testers we get, the better :-)
I'll look at
Michael Tautschnig wrote:
Could you give 4.0~beta2+experimental48 a try on this test system of yours? I've
already asked Mât and Nicolas to test this version as you might have just seen,
but the more testers we get, the better :-)
I'll look at it monday, I'll make a clone of our current syst
Michael Tautschnig wrote:
That might depend on when exactly you plan to do this migration.
We are planning on doing the migration when Squeeze is released,
whenever that is.
Worst case scenario is if we have to use some/various temporary storage
for backing the developers' /home, and obviously
Hi,
For the upcoming release of Squeeze we want to make a clean install on
some 400 or so developer desktops. Our issue is that we run software
RAID and LVM2 on those desktops and preservation of volumes in this
setup isn't supported by FAI at the moment. This is an issue since some
of the de
mamadou diop wrote:
Thank you Fredrik. Now i have understood your advice. But, why isn't
the hook called updatebase.SQUEEZE_X86 instead
of updatebase.DEFAULT?
updatebase.DEFAULT will run on all classes, you could have
updatebase.SQUEEZE_X86, updatebase.LENNY_AMD64 and so on but the command
is
t for each installation we use the method i
described in the previous message.
2010/11/23 Fredrik Eriksson <mailto:fredrik.eriks...@axis.com>>
mamadou diop wrote:
Because of the sources.list file which only points to a Debian
release at a given time, it seems to me
mamadou diop wrote:
Because of the sources.list file which only points to a Debian release
at a given time, it seems to me
that it is impossible to install simultaneously Debian clients with
differents releases. Is my opinion true?
We actually have this in our current testing environment.
What
job at our server.
Pbuilder defaults to cdebootstrap, which seems to be just debootstrap
written in C, but can easily be configured otherwise in /etc/pbuilderrc
(have a look at /etc/pbuilder/pbuilderrc, the default /etc/pbuilderrc
overrides, if you choose to install it.
Regards
--
Fredrik Eriksson
<>
26 matches
Mail list logo
