Ali Roze
2007-Feb-06 00:42 UTC
[Xen-users] xen bridged network config woes [repost w/apology]
My profuse apologies if you''ve already read this-it is a repost. I accidentally replied to a different thread, and I fear it was buried as a reply to that thread and may not get read by anyone but those who were paying attention to that one. My problem is unrelated to that one, other than that both involve xen''s networking, and I happened to be reading that one when I decided to write my own plea for help. Hi, I''ve been bashing my head over Xen networking for about a week now, and I''d love to get some help from the list. I''ve read the Wiki, the manual, the mailing list archives, and Googled my fingers to the bone. I''ll try to describe my problem as best as I can; if I''ve left anything relevant out, please let me know what you need. In a nutshell: -The server is at a colo facility, and I have no physical access to it, just ssh. -I have the IP addresses A.B.94.226 through 94.230. The gateway is at 94.225. -Dom0 is configured with A.B.94.226. I''ve only set up a single DomU, with A.B.94.227. I''m using Xen''s bridging scripts. -Dom0 can ping DomU. DomU can ping Dom0. Dom0 can reach the internet and be reached by it, but DomU cannot. DomU''s /etc/network/interfaces: root@domU:~# cat /etc/network/interfaces # Used by ifup(8) and ifdown(8). See the interfaces(5) manpage or # /usr/share/doc/ifupdown/examples for more information. # The loopback network interface auto lo iface lo inet loopback # The primary network interface # Uncomment this and configure after the system has booted for the first time auto eth0 iface eth0 inet static address A.B.94.227 netmask 255.255.255.248 gateway A.B.94.225 DomU config file: root@dom0:~# cat /etc/xen/domU # -*- mode: python; -*- kernel = "/boot/vmlinuz-2.6.16.29-xen" ramdisk = "/boot/initrd.img-2.6.16-29-xen" memory = 128 name = "domU" vif = [''bridge=xenbr0,ip=A.B.94.227''] disk [''phy:/dev/xenvolume/domU,sda1,w'',''phy:/dev/xenvolume/domU-swap,sda2,w''] netmask = "255.255.255.248" gateway = "A.B.94.225" hostname = "domU" root = "/dev/sda1 ro" Some more info from Dom0: root@dom0:~# brctl show bridge name bridge id STP enabled interfaces xenbr0 8000.feffffffffff no vif0.0 peth0 vif1.0 root@dom0:~# brctl showmacs xenbr0 port no mac addr is local? ageing timer 1 00:14:85:f7:ea:67 no 0.00 2 00:19:56:5a:e9:d5 no 1.03 2 00:60:3e:0b:9c:48 no 0.00 1 fe:ff:ff:ff:ff:ff yes 0.00 dom0 xend-config.sxp: root@dom0:~# cat /etc/xen/xend-config.sxp | grep -v "^#" | grep "[a-z]" (xend-relocation-server yes) (xend-relocation-hosts-allow ''^localhost$ ^localhost\\.localdomain$'') (network-script network-bridge) (vif-script vif-bridge) (dom0-min-mem 196) (dom0-cpus 0) I tried a tcpdump on dom0 while pinging an outside-the-network host from domU, here''s what I get: domU: root@domU:~# ping -c4 4.2.2.1 PING 4.2.2.1 (4.2.2.1) 56(84) bytes of data.>From A.B.94.227 icmp_seq=1 Destination Host Unreachable >From A.B.94.227 icmp_seq=2 Destination Host Unreachable >From A.B.94.227 icmp_seq=3 Destination Host Unreachable >From A.B.94.227 icmp_seq=4 Destination Host Unreachable--- 4.2.2.1 ping statistics --- 4 packets transmitted, 0 received, +4 errors, 100% packet loss, time 3002ms , pipe 3 root@domU:~# ping -c4 A.B.94.226 PING A.B.94.226 (A.B.94.226) 56(84) bytes of data. 64 bytes from A.B.94.226: icmp_seq=1 ttl=64 time=0.113 ms 64 bytes from A.B.94.226: icmp_seq=2 ttl=64 time=0.107 ms 64 bytes from A.B.94.226: icmp_seq=3 ttl=64 time=0.090 ms 64 bytes from A.B.94.226: icmp_seq=4 ttl=64 time=0.087 ms --- A.B.94.226 ping statistics --- 4 packets transmitted, 4 received, 0% packet loss, time 2997ms rtt min/avg/max/mdev = 0.087/0.099/0.113/0.013 ms Meanwhile in dom0: root@dom0:~# tcpdump src host A.B.94.227 -vv -a tcpdump: WARNING: vif0.0: no IPv4 address assigned tcpdump: listening on vif0.0, link-type EN10MB (Ethernet), capture size 96 bytes 14:57:36.016831 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:37.016795 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:38.016804 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:39.026781 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:40.026796 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:41.026795 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:54.872056 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 1, length 64 14:57:55.871054 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 2, length 64 14:57:56.870039 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 3, length 64 14:57:57.869040 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 4, length 64 14:57:59.864254 arp reply woc-2.ce.webhost.net is-at 00:16:3e:39:f0:ab (oui Unknown) 11 packets captured 25 packets received by filter 0 packets dropped by kernel root@dom0:~# As you can see, when domU is trying to ping 4.2.2.1, it''s sending ARP who-has requests that apparently never get answered. When domU is trying to ping dom0, it just pings it with no problems. So where should I start looking for a misconfiguration? How can I troubleshoot this? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Gary W. Smith
2007-Feb-06 00:58 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
Ali, What does the iptables scripts look like on the Dom0. I know that when I originally configured my Fedora6 server I forgot to disable it and a lot of things broke. Gary -----Original Message----- From: xen-users-bounces@lists.xensource.com [mailto:xen-users-bounces@lists.xensource.com] On Behalf Of Ali Roze Sent: Monday, February 05, 2007 4:42 PM To: xen-users@lists.xensource.com Subject: [Xen-users] xen bridged network config woes [repost w/apology] My profuse apologies if you''ve already read this-it is a repost. I accidentally replied to a different thread, and I fear it was buried as a reply to that thread and may not get read by anyone but those who were paying attention to that one. My problem is unrelated to that one, other than that both involve xen''s networking, and I happened to be reading that one when I decided to write my own plea for help. Hi, I''ve been bashing my head over Xen networking for about a week now, and I''d love to get some help from the list. I''ve read the Wiki, the manual, the mailing list archives, and Googled my fingers to the bone. I''ll try to describe my problem as best as I can; if I''ve left anything relevant out, please let me know what you need. In a nutshell: -The server is at a colo facility, and I have no physical access to it, just ssh. -I have the IP addresses A.B.94.226 through 94.230. The gateway is at 94.225. -Dom0 is configured with A.B.94.226. I''ve only set up a single DomU, with A.B.94.227. I''m using Xen''s bridging scripts. -Dom0 can ping DomU. DomU can ping Dom0. Dom0 can reach the internet and be reached by it, but DomU cannot. DomU''s /etc/network/interfaces: root@domU:~# cat /etc/network/interfaces # Used by ifup(8) and ifdown(8). See the interfaces(5) manpage or # /usr/share/doc/ifupdown/examples for more information. # The loopback network interface auto lo iface lo inet loopback # The primary network interface # Uncomment this and configure after the system has booted for the first time auto eth0 iface eth0 inet static address A.B.94.227 netmask 255.255.255.248 gateway A.B.94.225 DomU config file: root@dom0:~# cat /etc/xen/domU # -*- mode: python; -*- kernel = "/boot/vmlinuz-2.6.16.29-xen" ramdisk = "/boot/initrd.img-2.6.16-29-xen" memory = 128 name = "domU" vif = [''bridge=xenbr0,ip=A.B.94.227''] disk [''phy:/dev/xenvolume/domU,sda1,w'',''phy:/dev/xenvolume/domU-swap,sda2,w''] netmask = "255.255.255.248" gateway = "A.B.94.225" hostname = "domU" root = "/dev/sda1 ro" Some more info from Dom0: root@dom0:~# brctl show bridge name bridge id STP enabled interfaces xenbr0 8000.feffffffffff no vif0.0 peth0 vif1.0 root@dom0:~# brctl showmacs xenbr0 port no mac addr is local? ageing timer 1 00:14:85:f7:ea:67 no 0.00 2 00:19:56:5a:e9:d5 no 1.03 2 00:60:3e:0b:9c:48 no 0.00 1 fe:ff:ff:ff:ff:ff yes 0.00 dom0 xend-config.sxp: root@dom0:~# cat /etc/xen/xend-config.sxp | grep -v "^#" | grep "[a-z]" (xend-relocation-server yes) (xend-relocation-hosts-allow ''^localhost$ ^localhost\\.localdomain$'') (network-script network-bridge) (vif-script vif-bridge) (dom0-min-mem 196) (dom0-cpus 0) I tried a tcpdump on dom0 while pinging an outside-the-network host from domU, here''s what I get: domU: root@domU:~# ping -c4 4.2.2.1 PING 4.2.2.1 (4.2.2.1) 56(84) bytes of data.>From A.B.94.227 icmp_seq=1 Destination Host Unreachable >From A.B.94.227 icmp_seq=2 Destination Host Unreachable >From A.B.94.227 icmp_seq=3 Destination Host Unreachable >From A.B.94.227 icmp_seq=4 Destination Host Unreachable--- 4.2.2.1 ping statistics --- 4 packets transmitted, 0 received, +4 errors, 100% packet loss, time 3002ms , pipe 3 root@domU:~# ping -c4 A.B.94.226 PING A.B.94.226 (A.B.94.226) 56(84) bytes of data. 64 bytes from A.B.94.226: icmp_seq=1 ttl=64 time=0.113 ms 64 bytes from A.B.94.226: icmp_seq=2 ttl=64 time=0.107 ms 64 bytes from A.B.94.226: icmp_seq=3 ttl=64 time=0.090 ms 64 bytes from A.B.94.226: icmp_seq=4 ttl=64 time=0.087 ms --- A.B.94.226 ping statistics --- 4 packets transmitted, 4 received, 0% packet loss, time 2997ms rtt min/avg/max/mdev = 0.087/0.099/0.113/0.013 ms Meanwhile in dom0: root@dom0:~# tcpdump src host A.B.94.227 -vv -a tcpdump: WARNING: vif0.0: no IPv4 address assigned tcpdump: listening on vif0.0, link-type EN10MB (Ethernet), capture size 96 bytes 14:57:36.016831 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:37.016795 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:38.016804 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:39.026781 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:40.026796 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:41.026795 arp who-has woc.gw.webhost.net tell woc-2.ce.webhost.net 14:57:54.872056 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 1, length 64 14:57:55.871054 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 2, length 64 14:57:56.870039 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 3, length 64 14:57:57.869040 IP (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto: ICMP (1), length: 84) woc-2.ce.webhost.net > woc-1.ce.webhost.net: ICMP echo request, id 30474, seq 4, length 64 14:57:59.864254 arp reply woc-2.ce.webhost.net is-at 00:16:3e:39:f0:ab (oui Unknown) 11 packets captured 25 packets received by filter 0 packets dropped by kernel root@dom0:~# As you can see, when domU is trying to ping 4.2.2.1, it''s sending ARP who-has requests that apparently never get answered. When domU is trying to ping dom0, it just pings it with no problems. So where should I start looking for a misconfiguration? How can I troubleshoot this? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 01:06 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
On Mon, 5 Feb 2007 16:58:46 -0800, "Gary W. Smith" <gary@primeexalia.com> said:> What does the iptables scripts look like on the Dom0. I know that when > I originally configured my Fedora6 server I forgot to disable it and a > lot of things broke.Here''s my Dom0 iptables. However, I still have the same problem even if I drop all my iptables rules and leave the box wide open with iptables -F. root@Dom0:~# iptables -L Chain INPUT (policy ACCEPT) target prot opt source destination ACCEPT all -- anywhere anywhere state RELATED,ESTABLISHED ACCEPT tcp -- A.B.66.248/29 anywhere tcp dpt:ssh ACCEPT tcp -- owl.webhost.net anywhere tcp dpt:ssh ACCEPT tcp -- A.B.66.0/28 anywhere tcp dpt:ssh ACCEPT tcp -- A.B.68.38 anywhere tcp dpt:ssh ACCEPT tcp -- smtp.webhost.net anywhere tcp dpt:ssh ACCEPT all -- adsl-209-78-192-139.dsl.lsan03.pacbell.net anywhere ACCEPT tcp -- 70-32-242-119.ontrca.adelphia.net anywhere tcp dpt:ssh ACCEPT tcp -- A.B.66.248/29 anywhere tcp dpt:vmware-authd ACCEPT tcp -- 70-32-242-119.ontrca.adelphia.net anywhere tcp dpt:vmware-authd ACCEPT tcp -- anywhere anywhere tcp dpts:ftp-data:ftp ACCEPT tcp -- anywhere anywhere tcp dpt:smtp ACCEPT tcp -- anywhere anywhere tcp dpt:www ACCEPT tcp -- anywhere anywhere tcp dpt:pop3 ACCEPT tcp -- anywhere anywhere tcp dpt:ntp ACCEPT udp -- anywhere anywhere udp dpt:ntp DROP tcp -- anywhere anywhere tcp dpts:0:1023 DROP udp -- anywhere anywhere udp dpts:0:1023 ACCEPT tcp -- outgoing.webhost.net anywhere tcp dpts:amanda:10089 ACCEPT udp -- outgoing.webhost.net anywhere udp dpts:amanda:10089 ACCEPT tcp -- am.webhost.net anywhere tcp dpts:amanda:10089 ACCEPT udp -- am.webhost.net anywhere udp dpts:amanda:10089 ACCEPT tcp -- up.webhost.net anywhere tcp dpts:amanda:10089 ACCEPT udp -- up.webhost.net anywhere udp dpts:amanda:10089 DROP tcp -- anywhere anywhere tcp dpts:amanda:10089 DROP udp -- anywhere anywhere udp dpts:amanda:10089 Chain FORWARD (policy ACCEPT) target prot opt source destination ACCEPT all -- woc-2.ce.webhost.net anywhere PHYSDEV match --physdev-in vif1.0 ACCEPT udp -- anywhere anywhere PHYSDEV match --physdev-in vif1.0 udp spt:bootpc dpt:bootps Chain OUTPUT (policy ACCEPT) target prot opt source destination DROP icmp -- anywhere anywhere state INVALID _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Gary W. Smith
2007-Feb-06 01:21 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
> -----Original Message----- > From: xen-users-bounces@lists.xensource.com [mailto:xen-users- > bounces@lists.xensource.com] On Behalf Of Ali Roze > Sent: Monday, February 05, 2007 5:06 PM > To: Gary W. Smith; xen-users@lists.xensource.com > Subject: RE: [Xen-users] xen bridged network config woes [repost > w/apology] > > > Here''s my Dom0 iptables. However, I still have the same problem evenif> I drop all my iptables rules and leave the box wide open with iptables > -F. >Hmm. Not sure then. Did you make any custom changes to the bridge scripts? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 01:27 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
On Mon, 5 Feb 2007 17:21:13 -0800, "Gary W. Smith" <gary@primeexalia.com> said:> Hmm. Not sure then. Did you make any custom changes to the bridge > scripts?Not a thing! They''re the stock scripts that came with Xen (3.0.3). _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Curtis Doty
2007-Feb-06 02:42 UTC
Re: [Xen-users] xen bridged network config woes [repost w/apology]
6:42pm Ali Roze said:> > -The server is at a colo facility, and I have no physical access to it, > just ssh. > > -I have the IP addresses A.B.94.226 through 94.230. The gateway is at > 94.225.You''ve done a fine job of obfuscating your colo/isp identity.> root@dom0:~# brctl showmacs xenbr0 > port no mac addr is local? ageing timer > 1 00:14:85:f7:ea:67 no 0.00 > 2 00:19:56:5a:e9:d5 no 1.03 > 2 00:60:3e:0b:9c:48 no 0.00 > 1 fe:ff:ff:ff:ff:ff yes 0.00Two Cisco devices on the outside. Hrmm...I wonder if your colo/isp is running port security and only allows you one MAC for your one server? If so, you''ll have to route or NAT. Have you tried contacting your upstream provider? ../C _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 02:51 UTC
Re: [Xen-users] xen bridged network config woes [repost w/apology]
On Mon, 5 Feb 2007 18:42:30 -0800 (PST), "Curtis Doty" <Curtis@GreenKey.net> said:> You''ve done a fine job of obfuscating your colo/isp identity.Hi Curtis-I actually don''t care! The ISP is pixelgate.net, and the A.B part of my IP addresses is 66.254. I obfuscated that part because I saw someone else obfuscating theirs with A.B. while reading on the list, and I thought "neat idea, I''ll do that too". Monkey see, monkey do. Sorry. So my gateway is at 66.254.94.225.> Two Cisco devices on the outside. Hrmm...I wonder if your colo/isp is > running port security and only allows you one MAC for your one server? If > so, you''ll have to route or NAT. > > Have you tried contacting your upstream provider?I spent some time on the phone with a tech a few days ago, trying to figure out if anything looked misconfigured to him, but he said my domU''s /etc/interfaces and its xen config file looked fine to him. MAC addresses didn''t come up in our conversation at all. However, they don''t support Xen officially, and it may not have occurred to him to look for that. I''ve put in an email, asking them about that. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Gary W. Smith
2007-Feb-06 03:19 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
> I spent some time on the phone with a tech a few days ago, trying to > figure out if anything looked misconfigured to him, but he said my > domU''s /etc/interfaces and its xen config file looked fine to him. MAC > addresses didn''t come up in our conversation at all. However, theydon''t> support Xen officially, and it may not have occurred to him to lookfor> that. I''ve put in an email, asking them about that. >It really shouldn''t matter. When you are getting ranges of IP''s from your ISP they shouldn''t care weather or not the server behind it is a xen box or a physical box. If ISP''s start caring weather or not the server is a Xen box or not then we''re all in trouble. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Curtis Doty
2007-Feb-06 03:43 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
7:19pm Gary W. Smith said:> > I spent some time on the phone with a tech a few days ago, trying to > > figure out if anything looked misconfigured to him, but he said my > > domU''s /etc/interfaces and its xen config file looked fine to him. MAC > > addresses didn''t come up in our conversation at all. However, they > don''t > > support Xen officially, and it may not have occurred to him to look > for > > that. I''ve put in an email, asking them about that. > > > > It really shouldn''t matter. When you are getting ranges of IP''s from > your ISP they shouldn''t care weather or not the server behind it is a > xen box or a physical box. >Actually it can. If his upstream/ISP was designed by a Cisco engineer for single-server-per-port hosting, then it is conceivable they use "port security" to restrict access. Since Xen is a bridge by default and each DomU makes up MACs as needed, then they appear as multiple layer two hosts rather than just one. Thus my suggestion to try routing instead. Ali, are the other addresses in your /29 netblock also for use only on your one server? (vips) ../C _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 03:57 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
On Mon, 5 Feb 2007 19:43:17 -0800 (PST), "Curtis Doty" <Curtis@GreenKey.net> said:> Actually it can. If his upstream/ISP was designed by a Cisco engineer for > single-server-per-port hosting, then it is conceivable they use "port > security" to restrict access.I just heard back from them. The word is that if an unexpected MAC were to show up, a security measure in the switch would simply shut it down until someone from the ISP can look at it. Since that has never happened, by definition my domU''s packets have never even left the box and reached the switch! They were kind of amazed to find out from me that Xen makes up MAC addresses that may change for the same IP by default, and instructed me to define a single MAC address for each IP in my xen config, and give it to them so that they can tell the switch not to shut down.> suggestion to try routing instead.I don''t understand networking very well, which is why I just went with bridging because I understood from reading the Wiki that it was the It-Just-Works default. Aside from switching to the routing scripts in my xend-config.sxp, what do I need to do to make this work?> Ali, are the other addresses in your /29 netblock also for use only on > your one server? (vips)I''m not sure I understand your question, but I''ll try to answer it anyway. As configured by the host, the IPs were all assigned to my server, in the rc.local file. 94.226 is my main eth0 IP and the others are all aliases on eth0:227, eth0:228, eth0:229 and eth0:230. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 04:01 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
On Mon, 05 Feb 2007 21:57:27 -0600, "Ali Roze" <xen@ayahuasca.fastmail.us> said:> I''m not sure I understand your question, but I''ll try to answer it> anyway. As configured by the host, the IPs were all assigned to my > server, in the rc.local file. 94.226 is my main eth0 IP and the others > are all aliases on eth0:227, eth0:228, eth0:229 and eth0:230.I should hasten to add that I commented out the line for eth0:227 in dom0''s rc.local and restarted networking before setting up domU. If I don''t comment it out, I can''t even ping between domU and dom0, so I assume that''s the correct way to do it-assign the IP only in the domain in which it''s used. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Curtis Doty
2007-Feb-06 04:44 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
9:57pm Ali Roze said:> On Mon, 5 Feb 2007 19:43:17 -0800 (PST), "Curtis Doty" > <Curtis@GreenKey.net> said: > > Actually it can. If his upstream/ISP was designed by a Cisco engineer for > > single-server-per-port hosting, then it is conceivable they use "port > > security" to restrict access. > > I just heard back from them. The word is that if an unexpected MAC were > to show up, a security measure in the switch would simply shut it down > until someone from the ISP can look at it. Since that has never > happened, by definition my domU''s packets have never even left the box > and reached the switch! They were kind of amazed to find out from me > that Xen makes up MAC addresses that may change for the same IP by > default, and instructed me to define a single MAC address for each IP in > my xen config, and give it to them so that they can tell the switch not > to shut down. >Yep. It would appear that I guessed totally on target here. :-p Depending on the architecture, you may be able to get them to do this on your access interface: port security max-mac-count 2 Since it appears you are hosting all the other ip addresses on Dom0, you only need *one* additional DomU MAC (in addition to your existing Dom0 MAC) to swim up the bridge. This really isn''t an issue of security in the malicous "bad guy" sense. But more of a security design element that prevents catastrophe when some night-shift lackey plugs the wrong cable into the wrong port. :-/ (Never happened to me. Nope. Not once. Never. No way. Un uh. Impossible.)> > suggestion to try routing instead. > > I don''t understand networking very well, which is why I just went with > bridging because I understood from reading the Wiki that it was the > It-Just-Works default. Aside from switching to the routing scripts in my > xend-config.sxp, what do I need to do to make this work?This is actually adds complexity because, without re-design, it requires something called proxy ARP. I.e. your Dom0 presents its own layer two MAC on the wire so your other DomUs don''t have to. Plus it requires more netfilter tweaks.> > > Ali, are the other addresses in your /29 netblock also for use only on > > your one server? (vips) > > I''m not sure I understand your question, but I''ll try to answer it > anyway. As configured by the host, the IPs were all assigned to my > server, in the rc.local file. 94.226 is my main eth0 IP and the others > are all aliases on eth0:227, eth0:228, eth0:229 and eth0:230. >Yes you have answered it. Your upstream ISP is not Xen-aware, and is probably setup for a single layer two MAC per port. Hopefully you can now justify to them why you need just one more. ../C _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 05:06 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
Thank you Curtis-I''ve forwarded your email to them and I''m waiting for a reply. That said, while Googling, I found another interesting tcpdump incantation that I tried, and I''d appreciate some insight into what its output means: root@dom0:~# tcpdump -nvvi vif2.0 tcpdump: WARNING: vif2.0: no IPv4 address assigned tcpdump: listening on vif2.0, link-type EN10MB (Ethernet), capture size 96 bytes 21:01:36.829501 arp who-has 66.254.94.225 tell 66.254.94.227 21:01:37.215509 802.1d unknown version 21:01:37.829430 arp who-has 66.254.94.225 tell 66.254.94.227 21:01:39.230512 802.1d unknown version 21:01:39.619216 00:19:56:5a:e9:d5 > 01:00:0c:cc:cc:cc SNAP Unnumbered, ui, Flags [Command], length 89 21:01:39.829420 arp who-has 66.254.94.225 tell 66.254.94.227 21:01:40.829443 arp who-has 66.254.94.225 tell 66.254.94.227 21:01:41.246962 802.1d unknown version 8 packets captured 16 packets received by filter 0 packets dropped by kernel root@dom0:~# This raises three questions: 1. Do I care about that warning on the top, "vif2.0: no IPv4 address assigned"? 2. What does "802.1d unknown version" mean? 3. What does the "SNAP unnumbered" message mean? _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Curtis Doty
2007-Feb-06 05:33 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
11:06pm Ali Roze said:> > This raises three questions: > > 1. Do I care about that warning on the top, "vif2.0: no IPv4 address > assigned"? > 2. What does "802.1d unknown version" mean? > 3. What does the "SNAP unnumbered" message mean?This is getting a bit off-topic from Xen. The last two are well-established within the LAN (layer two) networking community. Do some sleuthing and enjoy the experience of learning. The first question, however, is related to Linux networking in general and Xen bridges specifically. But the answer is quite plain and apparent. It''s telling you that there is no IPv4 (layer three) address assigned to the interface (vif2.0) that you chose to monitor. That is all. And that is the nature of a bridge. It does not care about IP addresses. HTH, ../C _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Ali Roze
2007-Feb-06 05:40 UTC
RE: [Xen-users] xen bridged network config woes [repost w/apology]
On Mon, 5 Feb 2007 21:33:19 -0800 (PST), "Curtis Doty" <Curtis@GreenKey.net> said:> This is getting a bit off-topic from Xen. The last two are > well-established within the LAN (layer two) networking community. Do some > sleuthing and enjoy the experience of learning.That definitely helps, thank you. I just wanted to know whether these were clues to solving my xen problem or not. _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Curtis, you nailed it. That was exactly the problem and everything works great now. Thank you again! _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users