[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] problems with vlan+bonding


  • To: xen-users@xxxxxxxxxxxxxxxxxxx
  • From: Thomas van Oudenhove <thomasvo@xxxxxxxxxxxxxxxx>
  • Date: Fri, 26 Feb 2010 11:26:01 +0100
  • Delivery-date: Fri, 26 Feb 2010 02:27:05 -0800
  • List-id: Xen user discussion <xen-users.lists.xensource.com>
  • Openpgp: url=https://but.univ-toulouse.fr/pub/docs/ut.vanouden/public/ThomasvO_vanouden_AT_univ-toulouse.fr_0x26DF48EA_pub.asc

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

hi all,

I'm experiencing some problems on a domU with vlan and bonding.
sometimes, tcp packets are retransmitted, and it issues with a lag of
several seconds (4 to 10) on connexions (e.g. ssh) which makes it
unusable...

here is a small description of my stuff:

one dom0 (debian 5.0.4 (amd64), xen-hypervisor 3.2.1-2) with 4 eth
configured with bonding : eth0 and eth4 are enslaved in bond1, and eth1
and eth5 are enslaved in bond0. on bond1, I have a bridge with an IP
address for dom0, and my domUs are "plugged" on my private network
through it. on bond0, I have 2 VLANs bond0.90 and bond0.92, on each of
them a bridge, with no address for domO. the domUs public addresses are
"plugged" on those bridges.

here is my /etc/network/interfaces on dom0 (one of public VLAN/bridge is
not yet configured):
auto lo
iface lo inet loopback

# The primary bonding interface : private network
auto bond1
iface bond1 inet manual
        slaves eth0 eth4

auto brpriv
iface brpriv inet static
        bridge-ports bond1
        address 192.168.1.xxx
        netmask 255.255.255.0
        network 192.168.1.0
        broadcast 192.168.1.255
        dns-nameservers 192.168.1.xxx
        dns-search my-domain.com
        bridge_stp on
        bridge_fd 0

# the bonding interfaces for VLANs (802.1q)
auto bond0
iface bond0 inet manual
        slaves eth1 eth5

auto bond0.90
iface bond0.90 inet manual
        vlan_raw_device bond0

auto bond0.92
iface bond0.92 inet manual
        vlan_raw_device bond0

auto br92
iface br92 inet static
        bridge-ports bond0.92
        address 999.999.999.999
        netmask 255.255.255.000
        network 999.999.999.000
        broadcast 999.999.999.255
        gateway 999.999.999.1
        dns-nameservers 999.999.999.111 999.999.999.222
        dns-search my-domain.com
        bridge_stp on
        bridge_fd 0
# end of /etc/network/interfaces

here is my /etc/modprobe.d/bonding :
alias bond0 bonding
alias bond1 bonding
options bonding max_bonds=2 mode=802.3ad miimon=100 lacp_rate=1
use_carrier=0
# end of /etc/modprobe.d/bonding

I'm using the network-script dummy for setting domUs' network, and
vif-bridge for interfaces. my domUs are debian lennys (two 32 bits, and
one 64 bits).

with wireshark, I see at some time (seems to be random) retransmitted
TCP packets, which cause my ssh connexion to lag, and a TCP out-of-order
packet, and then everything comes back, until the next lag... these lags
only appear on public network, through the public switch and the tagged
VLAN (i.e. on private network, I don't see any problem; when I ssh from
one domU to another, I don't see any problem either, nor when I ssh from
dom0 to domU).

any help would be greatly appreciated to know where exactly the problem
is, and to solve.

thanks at least for reading ;),
regards,

- --
Thomas van Oudenhove - PRES Université de Toulouse
tél: (+33) 5 61 36 60 45
jabberID: thomasvo@xxxxxxxxxxxx
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkuHobUACgkQSvC+RibfSOrZ5gCfaLr75x0SCfKAOIn98ndChSCT
CXsAoOP3l+Muh6fP2L5MwHFXjsWbt6Ab
=o2PP
-----END PGP SIGNATURE-----

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.