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

Re: [Xen-users] newbie in trouble with CentOS Xen

  • To: xen-users@xxxxxxxxxxxxx
  • From: Alexandre Kouznetsov <alk@xxxxxxxxxx>
  • Date: Thu, 04 Apr 2013 12:16:15 -0600
  • Delivery-date: Thu, 04 Apr 2013 18:17:27 +0000
  • List-id: Xen user discussion <xen-users.lists.xen.org>


El 04/04/13 10:13, Dave Stevens escribió:
[root@skeena ~]# xm list
Name                                      ID Mem(MiB) VCPUs State   Time(s)
Domain-0                                   0     9879     8 r-----    870.8
bulkley                                    1     2048     4 -b----   3283.0
[root@skeena ~]#

[root@skeena ~]# brctl show
bridge name    bridge id        STP enabled    interfaces
virbr0        8000.000000000000    yes
xenbr0        8000.feffffffffff    no        vif1.0
xenbr1        8000.feffffffffff    no        vif1.1

This tells us what Domains have network interfaces in what bridges.
It seems like there is a mix of configurations. Looks fine, for a bridged networking (not NATted, see below), except there might be a problem with peth0 and peth1.

[root@skeena ~]# ifconfig
eth1      Link encap:Ethernet  HWaddr 00:30:48:CE:84:7F
           inet addr:  Bcast:  Mask:
           inet6 addr: fe80::230:48ff:fece:847f/64 Scope:Link
           RX packets:81102 errors:0 dropped:0 overruns:0 frame:0
           TX packets:34828 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:0
           RX bytes:10772813 (10.2 MiB)  TX bytes:2721383 (2.5 MiB)

lo        Link encap:Local Loopback
           inet addr:  Mask:
           inet6 addr: ::1/128 Scope:Host
           UP LOOPBACK RUNNING  MTU:16436  Metric:1
           RX packets:3086 errors:0 dropped:0 overruns:0 frame:0
           TX packets:3086 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:0
           RX bytes:4816900 (4.5 MiB)  TX bytes:4816900 (4.5 MiB)

peth0     Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
           UP BROADCAST NOARP  MTU:1500  Metric:1
           RX packets:0 errors:0 dropped:0 overruns:0 frame:0
           TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:0 (0.0 b)  TX bytes:0 (0.0 b)
           Interrupt:250 Base address:0xc000

peth1     Link encap:Ethernet  HWaddr FE:FF:FF:FF:FF:FF
           inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
           UP BROADCAST RUNNING NOARP  MTU:1500  Metric:1
           RX packets:42346 errors:0 dropped:0 overruns:0 frame:0
           TX packets:73652 errors:0 dropped:0 overruns:0 carrier:0
           collisions:0 txqueuelen:1000
           RX bytes:9145360 (8.7 MiB)  TX bytes:5504883 (5.2 MiB)
           Interrupt:251 Base address:0xe000
This is weired.

The common way to give network to DomUs is to define a bridge, put the physical interface into it permanently, and plug DomU's interfaces there ion demand (you probably found that in the Wiki). The very old network-bridge xen script was defining the bridges as following:
save IP configuration from eth0,
rename eth0 to peth0 (aka "physical eth0"),
create bridge called "eth0" (now that the name is free),
set IP configuration it took from the physical eth0 on the new bridge.

That way, before xend was brought up you had a plain eth0 to your network. After xend startup you had the same eth0 to your network. This proved to be little bit confusing and the configuration scripts produced unstable results, so late Xen documentation suggests to set up bridges using you platform means, and just tell Xen what bridge to use for what DomU's NIC.

Check this oldier wiki http://wiki.xensource.com/xenwiki/XenNetworking. It probably reflect better your scenario (and imho has nicer pictures)

What I see here are peth0 and peth1, they are supposed to be physical interfaces with a valid MAC, and be of the same nature. Actually, they have HWaddr FE:FF:FF:FF:FF:FF and one of them have a inet6 addr. eth1 seems to be a plain physical interface (I bet it's the one you use to connect to your server) while a bridge called "eth1" is expected. eth0 is simply missing.

Please tell me, how many physical network cards does your server have? How are they supposed to be used? Was any physical NIC added or removed recently? Maybe some NIC just died? The change of NIC number could confuse the network-script that badly. Did I told it was unstable?

It would be of great reference if you attach your DomU .cfg (under /etc/xen) file and /etc/xen/xend-config.sxp (I took the path form a Debian, it may be different in CentOS).

Please check what version of this components do you have installed:
Xen hypervisor
Linux kernel
xenutils (the name of package may be different in CentOS)
xentools (the name of package may be different in CentOS)

[root@skeena ~]# iptables -L -v
     0     0 ACCEPT     all  --  any    virbr0  anywhere    state RELATED,ESTABLISHED
     0     0 ACCEPT     all  --  virbr0 any
     0     0 ACCEPT     all  --  virbr0 virbr0  anywhere
     0     0 REJECT     all  --  any    virbr0  anywhere
anywhere            reject-with icmp-port-unreachable
     0     0 REJECT     all  --  virbr0 any     anywhere
anywhere            reject-with icmp-port-unreachable

This must be the rules to restrict traffic within a bridge, dedicated to internal communication between Dom0 and DomU's. Unused, apparently. The rest of ipfilter configuration is permissive, so it shall not be the issue.

As the general solution to your problem, in case you just want to "make it work back" and not doing any new implementation (Dom0 OS and Xe upgrade would be probably cleaner if reinstalled), I would suggest to re-do the networking configuration. Probably your DomU was configured with NATted networking, when it broke Dom0 kept it' sexternal IP address. The bridged networking is simpler and easer to debug.


Alexandre Kouznetsov

Xen-users mailing list



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