OpenVZ Forum


Home » General » Support » LVS and openvz
LVS and openvz [message #34245] Tue, 16 December 2008 15:13 Go to next message
ebasley is currently offline  ebasley
Messages: 14
Registered: February 2008
Junior Member
From: *fbx.proxad.net
got a huge problem with openvz and LVS.


take 2 nodes with Linux version 2.6.26-1-openvz-686 (Debian 2.6.26-11), one (vz1) has ldirectord on CT0 and one VE (ve1), next (vz2) has just one VE (ve2).
if ldirectord (from vz1) is directing traffic to vz2 : all work fine, but if traffic is redirected to local CTO (vz1) kernel thread ksoftirqd takes 100% of the cpu and nothing else happens !
With tcpdump we can see numerous paquets forwarded to ve1, something similar to a loop !!

setup :
-----

VZ1 :
----

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
link/ether 00:1f:29:0e:f9:4b brd ff:ff:ff:ff:ff:ff
inet6 fe80::21f:29ff:fe0e:f94b/64 scope link
valid_lft forever preferred_lft forever
3: br0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN
link/ether 00:1f:29:0e:f9:4b brd ff:ff:ff:ff:ff:ff
inet 192.168.2.22/24 brd 192.168.2.255 scope global br0
inet 192.168.2.206/32 scope global br0
inet6 fe80::21f:29ff:fe0e:f94b/64 scope link
valid_lft forever preferred_lft forever
11: veth213.0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN
link/ether 00:18:51:46:62:00 brd ff:ff:ff:ff:ff:ff
inet6 fe80::218:51ff:fe46:6200/64 scope link
valid_lft forever preferred_lft forever




/etc/network/interfaces :
auto br0
iface br0 inet static
address 192.168.2.22
netmask 255.255.255.0
network 192.168.2.0
broadcast 192.168.2.255
gateway 192.168.2.250
bridge_ports eth0

veth, eth0 are bridged in br0.
LVZ VIP (192.168.2.206) is supported by br0


ve1:
----

1: lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
inet 192.168.2.206/32 scope global lo
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever

3: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN
link/ether 00:18:51:59:12:98 brd ff:ff:ff:ff:ff:ff
inet 192.168.2.213/24 brd 192.168.2.255 scope global eth0
inet6 fe80::218:51ff:fe59:1298/64 scope link
valid_lft forever preferred_lft forever

/etc/network/interfaces :
auto lo
iface lo inet loopback
post-up /bin/ip addr add 192.168.2.206 dev lo


auto eth0
iface eth0 inet static
address 192.168.2.213
netmask 255.255.255.0
gateway 192.168.2.250
dns-nameservers 194.2.0.20 195.68.0.1
dns-search hdf-finance.fr
post-up /sbin/sysctl -w net.ipv4.conf.eth0.arp_ignore="1"
post-up /sbin/sysctl -w net.ipv4.conf.eth0.arp_annouce="2"

Help will be wellcome.
Thanks
Eric

Re: LVS and openvz [message #34259 is a reply to message #34245] Wed, 17 December 2008 18:27 Go to previous messageGo to next message
maratrus is currently offline  maratrus
Messages: 1495
Registered: August 2007
Location: Moscow
Senior Member
From: *sw.ru
Hello,

are you sure that it's OpenVZ specific problem?
Re: LVS and openvz [message #34268 is a reply to message #34259] Thu, 18 December 2008 16:03 Go to previous message
ebasley is currently offline  ebasley
Messages: 14
Registered: February 2008
Junior Member
From: 217.109.98*
may be not.

But the problem may come from the use of veth driver.
In all cases it would be very useful to have an HOWTO about using openvz in LVZ context.

Regards
Previous Topic: Openvz proxy arp or snat?
Next Topic: kernel problem !!
Goto Forum:
  


Current Time: Tue Jan 28 07:44:39 GMT 2020