FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian Kernel

 
 
LinkBack Thread Tools
 
Old 09-02-2008, 11:27 AM
Jan Grant
 
Default Bug#497528: openvz: Multiple bridges crashing when second veth device configured in the container.

Package: linux-image-2.6.26-1-openvz-amd64
Version: 2.6.26-4

I'm trying to get openvz working in a somewhat complex networking
environment, and hitting a hard lock-up when using multiple bridges.

I attach /etc/network/interfaces from CT0, plus the CT100 config and a
script used (manually at the moment) to bring up the bridges once the
CT100 has been started.

CT0 is networked to three separate subnets, called "main", "deploy" and
"ilom" respectively.

"deploy": this is physically plumbed into eth0.
"main": this is physically plumbed into bond0 (eth1 and eth2).
"ilom": this is physically plumbed into eth3.

There are three bridge devices created (and assigned IP addresses)
within CT0, named as above - see the attached interfaces for details.
The only subtle thing here is that the "main" bridge is on a bonded
interface.

This all works, and the box boots and can see its attached networks.

I can do the following:

vzctl start 100

and the zone comes up. The zone itself only has its ethIlom configured
and allocated a network address.

Once the vzbridge script (attached) is run, the veth devices of the new
CT100 are attached to their respective bridges.

All good so far: I have visibility to the ILOM network from CT0 and from
the VE.

However, once I attempt to bring up ethDeploy within the VE, the host
locks up hard:

vzctl enter 100
ifconfig ethDeploy up
[hard lockup]

I've a requirement to multiply-connect these containers, so any help
here would be appreciated. Certainly a hard lockup doesn't sound like
the correct behaviour.

Cheers,
jan


--
jan grant, ISYS, University of Bristol. http://www.bris.ac.uk/
Tel +44 (0)117 3317661 http://ioctl.org/jan/
ioctl(2): probably the coolest Unix system call in the world# The loopback network interface
auto lo
iface lo inet loopback

# The primary network interface
auto deploy
iface deploy inet static
bridge_ports eth0
bridge_maxwait 8
address 172.16.4.10
netmask 255.255.255.0
network 172.16.4.0
broadcast 172.16.4.255
# gateway 172.16.4.128
metric 20
# dns-* options are implemented by the resolvconf package, if installed
dns-nameservers 99.99.99.99
dns-search example.org
up route add -net 224.0.0.0/4 dev deploy
down route del -net 224.0.0.0/4 dev deploy

#auto bond0
iface bond0 inet manual
slaves eth1 eth2

auto main
iface main inet static
address 172.16.9.31
netmask 255.255.255.0
network 172.16.9.0
broadcast 172.16.9.255
gateway 172.16.9.250
metric 5
bridge_ports bond0
bridge_maxwait 8
pre-up /sbin/ifconfig bond0 up
pre-up /sbin/ifenslave bond0 eth1 eth2

auto ilom
iface ilom inet static
address 10.0.0.38
netmask 255.255.255.0
network 10.0.0.0
broadcast 10.0.0.255
metric 20
bridge_ports eth3
bridge_maxwait 8
#!/bin/sh
# /etc/vz/scripts/vzbridge
# a script to add virtual network interfaces (veth's) in a CT to a bridge on CT0

if [ -z "$VEID" ]; then
VEID="$1"
fi

CTID=$VEID

CONFIGFILE=/etc/vz/conf/$CTID.conf
. $CONFIGFILE

echo "$NETIF" | tr ';' '
' | while read line
do
HOSTIF=$( echo "$line" | tr ',' '
' | grep '^host_ifname=' | sed -e 's/^host_ifname=//' )
BRIDGE=$( echo "$line" | tr ',' '
' | grep '^bridge=' | sed -e 's/^bridge=//' )
if [ -n "$HOSTIF" -a -n "$BRIDGE" ]; then
echo "Adding interface $HOSTIF to bridge $BRIDGE on CT0 for CT$CTID"
/sbin/ifconfig $HOSTIF 0
echo 1 > /proc/sys/net/ipv4/conf/$HOSTIF/proxy_arp
echo 1 > /proc/sys/net/ipv4/conf/$HOSTIF/forwarding
/usr/sbin/brctl addif $BRIDGE $HOSTIF
fi
done

exit 0

KMEMSIZE="14372700:14790164"
LOCKEDPAGES="256:256"
PRIVVMPAGES="65536:69632"
SHMPAGES="21504:21504"
NUMPROC="240:240"
PHYSPAGES="0:9223372036854775807"
VMGUARPAGES="33792:9223372036854775807"
OOMGUARPAGES="26112:9223372036854775807"
NUMTCPSOCK="360:360"
NUMFLOCK="188:206"
NUMPTY="16:16"
NUMSIGINFO="256:256"
TCPSNDBUF="1720320:2703360"
TCPRCVBUF="1720320:2703360"
OTHERSOCKBUF="1126080:2097152"
DGRAMRCVBUF="262144:262144"
NUMOTHERSOCK="360:360"
NUMFILE="9312:9312"
DCACHESIZE="3409920:3624960"
NUMIPTENT="128:128"
AVNUMPROC="180:180"
CPUUNITS="1000"
ONBOOT="no"
ORIGIN_SAMPLE="vps.basic"
DISKSPACE="10485760:11530240"
DISKINODES="2000000:2200000"
QUOTATIME="0"
OSTEMPLATE=debian-4.0
NETIF="ifname=ethDeploy,mac=00:18:51:34:B3:C1,host _ifname=veth100.0,host_mac=00:18:51:13:2C:58,bridg e=deploy;ifname=ethIlom,mac=00:18:51:B6:34:B4,host _ifname=veth100.3,bridge=ilom,host_mac=00:18:51:FD :E6:FF;ifname=ethMain,mac=00:18:51:26:A5:1F,host_i fname=veth100.1,host_mac=00:18:51:5B:81:87"
# CONFIG_CUSTOMIZED="yes"
 
Old 09-02-2008, 12:19 PM
maximilian attems
 
Default Bug#497528: openvz: Multiple bridges crashing when second veth device configured in the container.

dear jan grant,

On Tue, Sep 02, 2008 at 12:27:08PM +0100, Jan Grant wrote:
> Package: linux-image-2.6.26-1-openvz-amd64
> Version: 2.6.26-4
>
> I'm trying to get openvz working in a somewhat complex networking
> environment, and hitting a hard lock-up when using multiple bridges.
>
> I attach /etc/network/interfaces from CT0, plus the CT100 config and a
> script used (manually at the moment) to bring up the bridges once the
> CT100 has been started.

i see, could you please file that upstream so that openvz devs read
about it and tell us the upstream bug number so that we can keep
track of it.
-> http://bugzilla.openvz.org/

> CT0 is networked to three separate subnets, called "main", "deploy" and
> "ilom" respectively.
>
> "deploy": this is physically plumbed into eth0.
> "main": this is physically plumbed into bond0 (eth1 and eth2).
> "ilom": this is physically plumbed into eth3.
>
> There are three bridge devices created (and assigned IP addresses)
> within CT0, named as above - see the attached interfaces for details.
> The only subtle thing here is that the "main" bridge is on a bonded
> interface.
>
> This all works, and the box boots and can see its attached networks.
>
> I can do the following:
>
> vzctl start 100
>
> and the zone comes up. The zone itself only has its ethIlom configured
> and allocated a network address.
>
> Once the vzbridge script (attached) is run, the veth devices of the new
> CT100 are attached to their respective bridges.
>
> All good so far: I have visibility to the ILOM network from CT0 and from
> the VE.
>
> However, once I attempt to bring up ethDeploy within the VE, the host
> locks up hard:
>
> vzctl enter 100
> ifconfig ethDeploy up
> [hard lockup]
>
> I've a requirement to multiply-connect these containers, so any help
> here would be appreciated. Certainly a hard lockup doesn't sound like
> the correct behaviour.
>
> Cheers,
> jan

thanks

--
maks



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 09-02-2008, 01:06 PM
Jan Grant
 
Default Bug#497528: openvz: Multiple bridges crashing when second veth device configured in the container.

On Tue, 2 Sep 2008, maximilian attems wrote:

> dear jan grant,

> i see, could you please file that upstream so that openvz devs read
> about it and tell us the upstream bug number so that we can keep
> track of it.
> -> http://bugzilla.openvz.org/

As requested, this is openvz #992:
http://bugzilla.openvz.org/show_bug.cgi?id=992

Many thanks.

--
jan grant, ISYS, University of Bristol. http://www.bris.ac.uk/
Tel +44 (0)117 3317661 http://ioctl.org/jan/
Donate a signature: http://ioctl.org/jan/sig-submit



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 09-03-2008, 09:10 PM
Jan Grant
 
Default Bug#497528: openvz: Multiple bridges crashing when second veth device configured in the container.

On Tue, 2 Sep 2008, Jan Grant wrote:

> As requested, this is openvz #992:
> http://bugzilla.openvz.org/show_bug.cgi?id=992
>
> Many thanks.

Those stirling chaps at openvz.org already have a patch that appears to
address this on lenny; can confirm it works for me with multiple VEs,
each of which has veth devices connected to multiple bridges.


--
jan grant, ISYS, University of Bristol. http://www.bris.ac.uk/
Tel +44 (0)117 3317661 http://ioctl.org/jan/
Unfortunately, I have a very good idea how fast my keys are moving.



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 

Thread Tools




All times are GMT. The time now is 08:38 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org