gmane.linux.ubuntu.server

Ubuntu Server Bug Triage Process

James Page — 2012-05-25T07:59:03

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi Team

Post discussion on this ML and at UDS-Q I've written up what I think
is the bug triage process we should be following for server bugs:

  https://wiki.ubuntu.com/ServerTeam/KnowledgeBase/BugTriage

I tested this during my triage session yesterday and it covered all of
the bugs that I triaged.

This is v1.0 - please review - happy to discuss changes and make
updates if people think that it needs improvement.

Cheers

James

- -- 
James Page
Ubuntu Core Developer
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQIcBAEBCAAGBQJPvzvHAAoJEL/srsug59jD+24QALaRwZZVmP+GFCaKv+yiY1TW
71BdbJT6nt3lWyVwKWk09JMe7mpSJ2/5ejgaGPdm21NFhGueCQNEWbFDHwDfMspv
iaZWVKQfg1+B1Awc8iSA0khm5CwqJb9wJvWWKQQVU6lhy+iOTcjYiwSVer1o+zIh
6/TY2qK//FSosSCDDzpP7sC46buYpgE/8e64/sqyJy8Yp8rz1/l6rsagyn+AcOMb
fi8qy3BnmPcKn/DGmtkUfOze5fyu2PcXnmPqz2Jm+11mLMxen48IBHMsVq9/tHjQ
F1cNXrBQs8NFTq403qz9KA0DeVJ3MiFHXy7gPT9+8KvW

Byobu autostart from cron

Whisperity — 2012-05-19T14:04:37

Dear fellow Ubuntu users,

I have an automatation question I'd like to ask from you. My server runs
some servers which are, unfortunately, not daemons (like Minecraft or
Source Dedicated Server). I am planning on making them automatically start
at boot with the *< at >reboot* directive in crontab, but I wish their sessions
to be usable from *byobu*.

Problem being, that byobu does not auto-start. I have written a shell
script:

#!/bin/bash
screen -dr byobu -X screen -t srcds
screen -r byobu -p srcds -X stuff "./start_gmod.sh $(printf '\r')"

And I execute it as my *nonprivileged* user in *crontab*:

< at >reboot           sourcesrv                    cd /srv/srcds/orangebox
&& ./hook-byobu.sh

However, running this script requires *byobu* to be started when it runs.
If I test the script with starting byobu by hand and then executing it from
SSH, it does order byobu to make a new window and then starts the server.

My question being: can I, somehow, automatize the start of byobu into the *
hook-byobu.sh* file? I tried

precise LIO target with iscsi support

Aleksi Handelofski — 2012-05-18T17:48:42

Hi All!

I saw that the kernel in precise brought support for the LIO target including the iscsi module.
So i wanted to know if the LIO iscsi support ist considered production ready or should we still use
the old and tested TGT?

Thanks for your opinions!

Alex

MAAS Install issue

Evan Jones — 2012-05-15T19:17:33

I have downloaded the newest Ubuntu server iso image and have problems with
setting up the cloud. Here is what has happened so far.

 

Using a VMWare 4.0 server as the install platform.

 

Attempt #1

.         created controller on VMWare server. Install went smooth with no
problems. IP address for MAAS controller was 192.168.1.9 (from network DHCP
server -Endian firewall).

.         Created maas superuser as per instructions. Worked correctly.

.         Opened IE9 on Win7 desktop and went to MAAS page 192.168.1.9/MAAS
as per instructions. Page opened correctly.

.         Performed sudo maas-import-isos. Worked correctly.

.         Created 2 nodes in VMWare. Installed correctly. Nodes shut down
correctly.

.         Opened IE9 on Win7 desktop and went to MAAS page 192.168.1.9/MAAS
as per instructions. Page opened correctly.

.         Attempted to "Accept and Commission" nodes. System hung on me. I
had to shutdown controller and reboot.

.         Could not get back to MAAS Admin page at 192.168.1.9/M

resolv.conf

Diác. C.J.Moretti — 2012-05-16T19:02:52

Hello!

I would put the DNS IP
resolv.conf file
But the reniciar the machine it loses this
setting ...
In which file and place this increase should I do?

Moretti
                                     \///
                                    (o o)
__________ oo0 - () - 0oo __________
Tarde te amei,
Beleza antiga e tão nova,
tarde te amei.

Estavas dentro de mim
e eu estava fora...
Estavas comigo
e eu não estava contigo...
______________________________
 (Sto. Agostinho)

Ubuntu 12.04 / kvm / libvirt / problem with network bridge br0

Thorsten Göllner — 2012-05-16T15:07:41

Ubuntu server 12.04 locate problems

Diác. C.J.Moretti — 2012-05-16T13:38:54

HI!

After installing ubuntu 4.12, I noticed that when using the command
man, I always get the following message, even showing me the correct 
syntax of the command you want:
man: can not set the locale, make sure $ LC_ * and $ LANG are correct

When enter the line:
export LANG = pt_BR

the problem does not happen ...

How can I fix this detail without having to enter every time the line up?

TKS
Moretti

Server-Installation 12.04 / automatic seciruty updates

Thorsten Göllner — 2012-05-16T13:29:06

Hi,

when installing ubuntu server 12.04 I was asked, of I want to active the 
automatic security update (service). I said "no". How can I install it 
now? Is it a specific apt-package?

Thanks for any hint
-Thorsten-

Use of Blueprints + Blueprint Template

James Page — 2012-05-16T08:34:29

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi Team

I've spent some time with various members of the Ubuntu Server Team
over the last few weeks reviewing how we use blueprints to track
progress during the development cycle.

This was really with the objective of trying to resolve a few
challenges that we have with current usage:

1) It's often hard to evaluate exactly when a blueprint has been
completed.

2) We sometimes have challenges with understanding the use cases for
work.

3) What the overall goal is for a blueprint is not always clear.

4) Endeavouring to bake QA into all work we do.

When I first started contributing to Ubuntu we did try to write
specifications associated with each blueprint (see [0] for the
template).

I think that this approach did not work that well - often the spec did
not get updated throughout the cycle as the primary focus is normally
on work item tracking in the blueprint - so they age and become pretty
useless very quickly.

That said there are a few concepts in the s

fsck of LVM snapshot - different behaviour in 12.04

Ryan Tandy — 2012-05-15T17:43:42

Hi list,

[I sent this a few days ago but it doesn't seem to have come through; 
apologies if you're reading it for a second time.]

I'd like to (ab)use LVM snapshots for a periodic online fsck in the 
background.  The essence of the script is something like this:

lvcreate -s -L $snapsize vg0/srv
e2fsck -p /dev/vg0/lvol0 # discard orphaned inodes
e2fsck -fy /dev/vg0/lvol0
lvremove -f vg0/lvol0

On Lucid, this seems to work well.  On Precise, I get some false 
positives almost every time I run it: wrong free blocks/inodes count, 
and/or bitmap differences.  The filesystem is actually OK; none of these 
minor inconsistencies show up during a boot-time fsck.

My understanding is that ext4 is wired up with LVM to do a sync and 
clean up these inconsistencies at the time a snapshot is made, and 
certainly on Lucid it seems OK.  Can anyone tell me if the behaviour has 
changed, or if I'm doing something wrong?

So far I've reproduced this on my workstation and laptop (one SATA drive 
each) and on one scratch serv

KVM network configuration issue

2012-05-10T14:35:34

Hi all.
I have a KVM guest webserver VWEB which I want to give an external IP.  My 
host has 4 NICs, of which I use only one presently. 

Host:
auto eth0
iface eth0 inet manual
auto br0
iface br0 inet static
address 172.16.5.11
netmask 255.255.0.0
gateway 172.16.1.1
dns-nameservers 172.16.1.2
bridge-ports eth0

I would like to connect a cable between my host's eth3 and my company FW. 
(That specific port of the FW is now in use by my old physical web server 
and all is working fine for this old server).
VWEB should use eth3 on the host.  And inside VWEB, i want to assign the 
fixed external IP.

I cannot find the doc for this setup.  I tried the following:
Host: i added the following:
auto br1
iface br1 inet manual
bridge-ports eth3

-> i didn't want my host to have this external ip available, only VWEB 
running as guest, so on the host I didn't configure the external IP.  I 
expected the packets form VWEB to be just passed to the interface.  It 
didn't work.  I have some ideas about why it may not work, but

cobbler how to set static ip address for ubuntu

Shake Chen — 2012-05-09T12:47:11

Hi

I have set up orchestra in ubuntu 12.04. use cobbler deploy OS

the Centos and ubuntu working well .

Now I need set the static IP address for the node.

In centos , I have success,

but for Ubuntu, it no working.

I read http://terrarum.net/administration/deploying-ubuntu-with-cobbler.html

can not working.

Ubuntu 12.04 / kvm / convirt can not be installed properly

Thorsten Göllner — 2012-05-08T07:53:06

lxc ubuntu-cloud template issue

Aljoša Mohorović — 2012-05-08T01:27:02

i'm getting:

No config file specified, using the default config
/usr/lib/lxc/templates/lxc-ubuntu-cloud: line 240: type:
ubuntu-cloudimg-query: not found
failed to execute template 'ubuntu-cloud'
aborted

when i use "-t ubuntu-cloud" but every doc i've read uses ubuntu-cloud
and doesn't mention my issue.
any ideas?

Aljosa Mohorovic

MaaS/Juju questions

Anders Bruun Olsen — 2012-05-07T13:14:07

Hi,

I am trying out MaaS and Juju for the first time and have run into a couple
of questions that I haven't been able to find answers to in the
documentation or by searching.

My setup is as follows: One server installed from scratch with the "create
maas server" option on the 12.04 server cdrom and two nodes that PXE boot
from the MaaS server.

1. I have the two nodes able to be started from the MaaS webinterface with
WoL. However, once they are booted I have been unable to find a way of
shutting them down or rebooting them except by physically walking up to
them and pressing the power button or ctrl-alt-delete. Is there no remote
way of doing this?

2. In continuance of Q1, I have added my SSH public key in the MaaS
webinterface, but I have been unable to login to any of the nodes over SSH.
Are there no accounts enabled on MaaS nodes? (this would enable remote
shutdown/reboot)

3. I have setup ~/.juju/environments.yaml according to the MaaS guide with
the OAuth key for my MaaS user. When I run 'juju boots

Ubuntu 12.04, KVM: STP and VLAN filtering from libvirt?

Andreas Ntaflos — 2012-05-04T00:33:16

Hi,

Reading the release notes for the server edition of 12.04, it says "STP
and VLAN filtering from libvirt" for the KVM/Libvirt changes.

What does this mean and where can I read more about it?

Thanks,

Andreas

UDS-Q Topic: Server Bug Triage Process

James Page — 2012-05-03T15:23:41

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi All

At UDS-Q we will be discussing the relatives merits and deficiencies
of the bug triage process that we have been using in the server team
for the last 12 months or so (see [0]) as well as decided on how to
approach the current backlog of bugs.

FYI people on the weekly triage rota normally use the triage report
[1] to identify bugs to work on for their allocated slots.

I'd be interested in hearing peoples opinions on the current process
in advance of UDS - specifically what you think works and what does
not and how you think things might be improved.

Regards

James


[0] https://wiki.ubuntu.com/ServerTeam/KnowledgeBase#triager
[1] http://status.qa.ubuntu.com/reports/ubuntu-server/triage-report.html
- -- 
James Page
Ubuntu Core Developer
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQIcBAEBCAAGBQJPoqL9AAoJEL/srsug59jD4wQP/AwDMbVNoSf3GYNFVnOwvwrf
IClUAyStGc9vwTXjtsCiOR

upgraded to precise, doesn't seem to shutdown

Craig White — 2012-04-30T21:33:39

Upgraded one our lightly used test server from Lucid to Precise and it seems that I can't get it to restart without pulling the power.

via a remote ssh shell, I instructed it to restart ( shutdown now -r) and syslog shows...

Apr 30 14:02:16 nxpc kernel: [269718.509842] init: tty4 main process (1104) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.510137] init: tty5 main process (1116) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.510429] init: tty2 main process (1142) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.510709] init: tty3 main process (1144) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.510998] init: tty6 main process (1146) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.511804] init: cron main process (1155) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.515293] init: irqbalance main process (1183) killed by TERM signal
Apr 30 14:02:16 nxpc kernel: [269718.516907] init: tty1 main process (1824) killed by TERM signal
Apr 30

OpenMPI for Q

Robie Basak — 2012-04-30T11:39:53

Dear List,

I have been asked whether we need to do anything for OpenMPI for Q. I
had a blueprint for P on this, for which I ended up delivering a 1.5
package in the archive.

Current status to the best of my knowledge is:

  * Upstream use an old Linux-style odd/even beta/stable version
    numbering scheme.

  * 1.4 is the current stable/production upstream release, is in Debian
    and Ubuntu and is the default.

  * 1.5 is the current upstream beta/testing release. Debian have it in
    experimental as replacing 1.4, but don't intend to upload it to sid
    until upstream declare it stable/production at which point it will be
    labelled 1.6. Ubuntu now has it as a separate package that can be
    used in parallel with 1.4 and built against by hand. But all
    packages in the Ubuntu archive that use OpenMPI are built against
    1.4 (and depend on the 1.4 package).

  * Both 1.4 and 1.5 have some level of ARM support. 1.4's ARM support
    is based on Jani's patches which we (Ubuntu) are carrying. 1.5'

Mysql service does not restart after unattended upgrade on Hardy

Lorenzo Salvadorini — 2012-04-27T10:45:14

Hi all,
we are experiencing this issue on many of our hardy hosts with
unattended upgrades enabled.
After upgrade from 5.0.51 to version 5.0.95 or from 5.0.95 to 5.0.96
the upgrade fails and the mysql service does not restarts correctly.
With a manual restart the mysql service come up without any apparent
problem but the version is still the old one. The day after, we have
done the upgrade manually without any problem. We experienced this
problem on different platform and setup:
- on EC2 hosts and on physical hosts on common HP DL360 G6/G7 hardware.
- on mysql clustered with DRBD and standalone hosts

I'm trying to understand what could be causing this behaviour reading
changelog or unattended upgrade logs/mail but I've not found anything
until now, is this happened to someone else?


More info about our typical setup affected by the issue:

~# lsb_release -r
Release:        8.04

~# uname -a
Linux foo02 2.6.24-28-server #1 SMP Sat Oct 16 18:09:42 UTC 2010
x86_64 GNU/Linux

Aptitude log:
/var/log/aptitu

Setgid not preserved in GFS2 with ACL

Nicolas Ecarnot — 2012-04-25T12:35:49

[Sorry for cross-posting, but I sincerely don't know who's best to answer]

Hi,

Using many production Samba file servers on RHEL 5.6 for a while, we are 
now finishing to setup a samba cluster on Ubuntu-server (oneiric) with 
cman+clvm+GFS2+ctdb.

Like on our other samba setups, we are using ACLs and we set up the 
setgid bit on our folders (chmod g+s folder), as well as default ACL.
The access rights are managed via the basic windows explorer security 
tab and is working nicely.

But on this new GFS2, I observe that this is not working the same.

To make it short, the setgid bit gets lost when a user creates a subdir.

To be precise, here is what I'm observing :

My folder looks like this :

root< at >server:/foo/bar# getfacl .
# file: .
# owner: root
# group: adminsGroup
# flags: ss-
user::rwx
group::rwx
group:domainUsers:rwx
mask::rwx
other::---
default:user::rwx
default:group::rwx
default:group:domainUsers:rwx
default:mask::rwx
default:other::---

* When the user root runs 'mkdir rootDir', this directory cor