Ubuntu
lxc package

corrupted kernel messages being logged to syslog

Bug #1003888 reported by James Hunt
36
This bug affects 8 people
Affects Status Importance Assigned to Milestone
lxc (Ubuntu)
Confirmed
High
Unassigned

Bug Description

Running a i686 32-bit quantal container on a i686 32-bit precise host I'm seeing kernel message corruption in /var/log/syslog in the container:

___________
May 24 10:41:04 localhost ntpdate[172]: adjust time server 91.189.94.4 offset 0.
001225 sec
May 24 10:41:05 localhost kernel: [4.531 t0 oIv otr rsn
___________

May 24 10:44:06 localhost ntpdate[199]: adjust time server 91.189.94.4 offset -0
.000260 sec
May 24 10:44:08 localhost kernel: [11726.678703] vethKvdkPd: no IPv6 routers pre
sent
May 24 10:44:08 localhost kernel: [11726.934078] eth0: no IPv6 routers present
May 24 10:44:26 localhost kernel: <>[1174.942 ye10 ui(37526584) pamr"EID prto=mu
t no"aldfasmth ro=1 aet198poie"x-otie-eal"nm=/a/i/raaeddbgs"pd105cm=uedha"ftp=db
gs rnm=nn"fas"w
May 24 10:45:31 localhost kernel: .33]eyL:oP urpst<6>[11809.800093] ADDRCONF(NET
DEV_CHANGE): veth3jQTu8: link becomes ready
May 24 10:45:31 localhost kernel: [11809.800220] lxcbr0: port 1(veth3jQTu8) ente
ring forwarding state
May 24 10:45:31 localhost kernel: [11809.800265] lxcbr0: port 1(veth3jQTu8) ente
ring forwarding state
May 24 10:45:42 localhost kernel: [11820.506530] et0:n P6otsre
May 24 10:45:46 localhost kernel: 684773]lcr:pr (ehjT8 neigfradn tt
May 24 10:57:11 localhost james: hello from quantal lxc
___________

The host has 2 examples:

May 24 12:02:20 azul kernel: [11811.087590] type=1400 audit(1337856332.979:49): apparmor="DENIED" operation="moun" if=fie lg ac"err-3prn=97 rfl=lccnanrdfut ae"vrlbuedha/euf/ i=99 om"raaed sye"euf"scae"oe lg=r"h oIvrurpst<6>[12815.645687] device vethemBouM entered promiscuous mode
May 24 12:02:20 azul kernel: [12815.64492]ADCN(EDVU) ehmoM iki o ed

___________
May 24 12:02:24 azul ntpd[3197]: new interface(s) found: waking up resolver
May 24 12:02:31 azul kernel: -fut ae"vrlbuedha/euf/ i=47 om"raaed sye"euf"scae"oe lg=r"ere
May 24 12:05:01 azul CRON[25554]: (root) CMD (command -v debian-sa1 > /dev/null && debian-sa1 1 1)
___________

Kernel version:

$ uname -a
Linux azul 3.2.0-24-generic-pae #38-Ubuntu SMP Tue May 1 16:40:26 UTC 2012 i686 i686 i386 GNU/Linux
$

ProblemType: Bug
DistroRelease: Ubuntu 12.04
Package: lxc 0.7.5-3ubuntu56
ProcVersionSignature: Ubuntu 3.2.0-24.38-generic-pae 3.2.16
Uname: Linux 3.2.0-24-generic-pae i686
ApportVersion: 2.0.1-0ubuntu7
Architecture: i386
Date: Thu May 24 12:03:41 2012
InstallationMedia: Ubuntu 10.10 "Maverick Meerkat" - Release i386 (20101007)
ProcEnviron:
 TERM=xterm
 PATH=(custom, user)
 LANG=fr_CA.UTF8
 SHELL=/bin/bash
SourcePackage: lxc
UpgradeStatus: Upgraded to precise on 2012-01-12 (132 days ago)

Revision history for this message
James Hunt (jamesodhunt) wrote :
Revision history for this message
Serge Hallyn (serge-hallyn) wrote :

Thanks James.

Do you see any corruption in the host's /var/log/syslog, or only in the container?

Changed in lxc (Ubuntu):
status: New → Incomplete
Revision history for this message
James Hunt (jamesodhunt) wrote :

Hi Serge - two corruptions in the host as mentioned above.

Revision history for this message
Serge Hallyn (serge-hallyn) wrote :

Thanks. The syslog namespace, which is not yet implemented, would be the one to help prevent this.

Can you try adding

syslog

to the line

lxc.cap.drop = sys_module mac_admin

in the container's configuration file (/var/lib/lxc/<container>/config)? So the line would then read:

lxc.cap.drop = sys_module mac_admin syslog

Does the container then start all right, and leave the host's syslog uncorrupted?

Changed in lxc (Ubuntu):
status: Incomplete → Confirmed
importance: Undecided → High
Revision history for this message
Alan AZZERA (azzera-alan) wrote :

Hi,

I saw that this bug was marked as duplicate of bug #460925. Anyhow, I wanted to report that under Precise I'm facing the same problem. The host's kern.log and syslog are clean, until I start a container. Then I can read in kern.log things like :
May 23 09:20:58 <$host> kernel: [ 4715.831085] EXT4-fs (dm-2): mounted filesystem with ordered data mode. Opts: (null)
May 23 09:29:35 <$host> kernel: [ 5233.548856] device veth_openm entered promiscuous mode
May 23 09:29:35 <$host> kernel: [ 5233.549397] ADDRCONF(NETDEV_UP): veth_openm: link is not ready
May 23 09:29:35 <$host> kernel: [ 5233.709180] ADDRCONF(NETDEV_CHANGE): veth_openm: link becomes ready
May 23 09:29:35 <$host> kernel: [ 5233.709236] br0: port 2(veth_openm) entering forwarding state
May 23 09:29:35 <$host> kernel: [ 5233.709241] br0: port 2(veth_openm) entering forwarding state
May 23 09:29:35 <$host> kernel: [ 5233.722420] audit_printk_skb: 3 callbacks suppressed
May 23 09:29:35 <$host> kernel: [ 5233.722425] type=1400 audit(1369294175.991:13): apparmor="DENIED" operation="mount" info="failed flags match" error=-13 parent=2799 profile="lxc-container-default" name="/var/lib/ureadahead/debugfs/" pid=2848 comm="ureadahead" fstype="debugfs" srcname="none" flags="rw"
May 23 09:29:46 <$host> kernel: <>[ 24000]vt_pn:n P6ruespeet<7[54.806 t0 oIv otr rsn
May 23 09:29:51 <$host> kernel: [ 5248760]b0 ot2vt_pn)etrn owrn tt

And in syslog :
May 23 09:29:35 <$host> kernel: [ 5233.709241] br0: port 2(veth_openm) entering forwarding state
May 23 09:29:35 <$host> kernel: [ 5233.722420] audit_printk_skb: 3 callbacks suppressed
May 23 09:29:35 <$host> kernel: [ 5233.722425] type=1400 audit(1369294175.991:13): apparmor="DENIED" operation="mount" info="failed flags match" error=-13 parent=2799 profile="lxc-container-default" name="/var/lib/ureadahead/debugfs/" pid=2848 comm="ureadahead" fstype="debugfs" srcname="none" flags="rw"
May 23 09:29:46 <$host> kernel: <>[ 24000]vt_pn:n P6ruespeet<7[54.806 t0 oIv otr rsn
May 23 09:29:51 <$host> kernel: [ 5248760]b0 ot2vt_pn)etrn owrn tt
May 23 10:17:01 <$host> CRON[3471]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly)

My Ubuntu server 12.04 is up to date and its kernel is 3.2.0-43-generic x86_64.

I also tried to modify a container config following the suggested way :
lxc.cap.drop = sys_module mac_admin syslog

The container does not start. That is the reason why I wrote here.

Best regards,

Al

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.
Subscribing...

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.