Ian Pratt
2005-Sep-14 20:49 UTC
RE: [Xen-users] veth0 is from netback and vifu.0 is fromthebridge-utils?
> > It''s effectively a point to point link, allowing domain 0 > to connect > > on to the bridge in the same manner that other domains do. (before, > > packets to domain 0 were short-cutted and came off the bridge > > directly, which could lead to traffic from other domains > being stalled > > if a lot of traffic was destined to user space in dom0. > > thx. what''s the diff betwn the two (veth0 and vifu.0) ?vifX.0 is the end of the virtual ''cross-over cable'' that''s connected to the bridge. veth0 is the end that''s effectively the domain''s virtual NIC, and hence has the IP address associated with it. Ian _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Anthony.Golia@MorganStanley.com
2005-Oct-24 17:56 UTC
RE: [Xen-users] veth0 is from netback and vifu.0 is fromthebridge-utils?
sorry to bring this up again but i''m still confused (i know things
changed
recently with peth0 and all). so theres no more veth0? whats peth0?
(someone said it was the physical int, i thought that was eth0, eth1,
etc.). Also, i''m using vif-route not bridge:
bash-3.00# ip addr
1: peth0: <BROADCAST,MULTICAST,NOARP,UP> mtu 1500 qdisc pfifo_fast qlen
1000
link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff
2: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000
link/ether 00:09:6b:b5:6b:95 brd ff:ff:ff:ff:ff:ff
inet 172.31.215.177/24 brd 172.31.215.255 scope global eth1
3: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue
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 172.31.254.207/32 brd 172.31.254.207 scope global lo:hostname
4: vif0.0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff
5: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
link/ether 00:09:6b:b5:6b:94 brd ff:ff:ff:ff:ff:ff
inet 172.31.205.177/24 brd 172.31.205.255 scope global eth0
6: xen-br0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff
7: vif1.1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue
link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff
inet 169.254.1.0/32 brd 169.254.1.0 scope global vif1.1
On Wed, 14 Sep 2005, Ian Pratt wrote:
>
> > > It''s effectively a point to point link, allowing domain
0
> > to connect
> > > on to the bridge in the same manner that other domains do.
(before,
> > > packets to domain 0 were short-cutted and came off the bridge
> > > directly, which could lead to traffic from other domains
> > being stalled
> > > if a lot of traffic was destined to user space in dom0.
> >
> > thx. what''s the diff betwn the two (veth0 and vifu.0) ?
>
> vifX.0 is the end of the virtual ''cross-over cable''
that''s connected to
> the bridge. veth0 is the end that''s effectively the
domain''s virtual
> NIC, and hence has the IP address associated with it.
>
> Ian
>
> )
>
Cheers,
Anthony
_______________________________________________
Xen-users mailing list
Xen-users@lists.xensource.com
http://lists.xensource.com/xen-users
Nivedita Singhvi
2005-Oct-24 19:57 UTC
Re: [Xen-users] veth0 is from netback and vifu.0 is fromthebridge-utils?
Anthony.Golia@MorganStanley.com wrote:> sorry to bring this up again but i''m still confused (i know things changed > recently with peth0 and all). so theres no more veth0? whats peth0? > (someone said it was the physical int, i thought that was eth0, eth1, > etc.). Also, i''m using vif-route not bridge: > > bash-3.00# ip addr > 1: peth0: <BROADCAST,MULTICAST,NOARP,UP> mtu 1500 qdisc pfifo_fast qlen > 1000 > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > 2: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 > link/ether 00:09:6b:b5:6b:95 brd ff:ff:ff:ff:ff:ff > inet 172.31.215.177/24 brd 172.31.215.255 scope global eth1 > 3: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue > 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 172.31.254.207/32 brd 172.31.254.207 scope global lo:hostname > 4: vif0.0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > 5: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > link/ether 00:09:6b:b5:6b:94 brd ff:ff:ff:ff:ff:ff > inet 172.31.205.177/24 brd 172.31.205.255 scope global eth0 > 6: xen-br0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > 7: vif1.1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > inet 169.254.1.0/32 brd 169.254.1.0 scope global vif1.1 >Anthony, what version/changeset are you running? Any answer you get is probably true only for a range of changesets. We''re trying to put together some documentation of the networking stack this week, once we clean up a few issues... Firstly, you seem to be running in bridged mode - you have all the infrastructure we create in the default bridge scripts (which we don''t in the default routing setup): The 0 in the names below is the default, it will be $vifnum, specific to your local script changes, otherwise... Domain0 network stack ^^ || || eth0 [local virtual interface] ^^ || || local traffic intended for domain0 || || xen-br0 [bridge] <===> vif0.0 [virtual dom0] <=======> eth0 [domU virtual nic] || || || external traffic || VV peth0 [real, physical interface] || --XX ------------ NETWORK ----------------- I''m trying to put together some documentation of the networking stack this week, once a few issues get cleaned up. I started on this lastweek but the stack has been a bit volatile (someone else floated something similar more recently with some more text).. Hope that helps for now.. thanks, Nivedita _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Anthony.Golia@MorganStanley.com
2005-Oct-24 20:23 UTC
Re: [Xen-users] veth0 is from netback and vifu.0 is fromthebridge-utils?
On Mon, 24 Oct 2005, Nivedita Singhvi wrote:> Anthony.Golia@MorganStanley.com wrote: > > sorry to bring this up again but i''m still confused (i know things changed > > recently with peth0 and all). so theres no more veth0? whats peth0? > > (someone said it was the physical int, i thought that was eth0, eth1, > > etc.). Also, i''m using vif-route not bridge: > > > > bash-3.00# ip addr > > 1: peth0: <BROADCAST,MULTICAST,NOARP,UP> mtu 1500 qdisc pfifo_fast qlen > > 1000 > > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > > 2: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 > > link/ether 00:09:6b:b5:6b:95 brd ff:ff:ff:ff:ff:ff > > inet 172.31.215.177/24 brd 172.31.215.255 scope global eth1 > > 3: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue > > 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 172.31.254.207/32 brd 172.31.254.207 scope global lo:hostname > > 4: vif0.0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > > 5: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > > link/ether 00:09:6b:b5:6b:94 brd ff:ff:ff:ff:ff:ff > > inet 172.31.205.177/24 brd 172.31.205.255 scope global eth0 > > 6: xen-br0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > > 7: vif1.1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > > inet 169.254.1.0/32 brd 169.254.1.0 scope global vif1.1 > > > > Anthony, what version/changeset are you running? > Any answer you get is probably true only for a range of changesets. > > We''re trying to put together some documentation of the networking > stack this week, once we clean up a few issues... > > Firstly, you seem to be running in bridged mode - you have all the > infrastructure we create in the default bridge scripts (which we > don''t in the default routing setup): > > The 0 in the names below is the default, it will be $vifnum, > specific to your local script changes, otherwise... > > > Domain0 network stack > ^^ > || > || > eth0 [local virtual interface] > ^^ > || > || local traffic intended for domain0 > || > || > xen-br0 [bridge] <===> vif0.0 [virtual dom0] <=======> eth0 [domU virtual nic] > || > || > || external traffic > || > VV > peth0 [real, physical interface] > || > --XX ------------ NETWORK ----------------- > > > I''m trying to put together some documentation of the networking > stack this week, once a few issues get cleaned up. I started on > this lastweek but the stack has been a bit volatile (someone else > floated something similar more recently with some more text).. > > Hope that helps for now.. > > thanks, > Nivedita >thx. that helps but leaves me with a few more questions. any doc is appreciated. i had bridging on then changed the xend-config.sxp to use vif-route so now i have a completely different set of ints (see below). i thought the eth0 in domU was linked to vifx.U in dom0? so with routing on there''s no peth0? is eth0 and eth1 still my real, physical interface? the ascii diagram you have above is helpfull how does it look with routing? bash-3.00# ip addr 1: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:09:6b:b5:6b:94 brd ff:ff:ff:ff:ff:ff inet 172.31.205.177/24 brd 172.31.205.255 scope global eth0 2: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 link/ether 00:09:6b:b5:6b:95 brd ff:ff:ff:ff:ff:ff inet 172.31.215.177/24 brd 172.31.215.255 scope global eth1 3: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue 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 172.31.254.207/32 brd 172.31.254.207 scope global lo:hostname 4: vif0.0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff 5: veth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff 6: vif1.1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff inet 169.254.1.0/32 brd 169.254.1.0 scope global vif1.1 Cheers, Anthony _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Nivedita Singhvi
2005-Oct-24 21:08 UTC
Re: [Xen-users] veth0 is from netback and vifu.0 is fromthebridge-utils?
Anthony.Golia@MorganStanley.com wrote:> thx. that helps but leaves me with a few more questions. any doc is > appreciated. i had bridging on then changed the xend-config.sxp > to use vif-route so now i have a completely different set of ints > (see below). i thought the eth0 in domU was linked to vifx.U in dom0? > so with routing on there''s no peth0? is eth0 and eth1 still my real, > physical interface? the ascii diagram you have above is helpfull how does > it look with routing? > > bash-3.00# ip addr > 1: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 > link/ether 00:09:6b:b5:6b:94 brd ff:ff:ff:ff:ff:ff > inet 172.31.205.177/24 brd 172.31.205.255 scope global eth0 > 2: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 > link/ether 00:09:6b:b5:6b:95 brd ff:ff:ff:ff:ff:ff > inet 172.31.215.177/24 brd 172.31.215.255 scope global eth1 > 3: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue > 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 172.31.254.207/32 brd 172.31.254.207 scope global lo:hostname > 4: vif0.0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > 5: veth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop > link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff > 6: vif1.1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > inet 169.254.1.0/32 brd 169.254.1.0 scope global vif1.1This really depends on your patchset/version you are running, and your local script. Are you running the default network-route and vif-route that''s currently in the unstable tree? If you change the vif file to be vif-route, you should have changed the network file to be network-route, too. Looks like you have two interfaces in dom0, that''s why you have both eth0 and eth1. You have two virtual backends (vif0.0, vif1.1). The default route scripts don''t rename the interface (so you won''t have a peth0, for instance).. Is the above working for you? Can you send netstat -rn output? In brief: physical interface (e.g. eth0) <=> virtual backend (e.g. vif0.0) || \\ =====> domU nic (eth0) I''ll confirm this once I get the latest tip working... thanks, Nivedita _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users
Anthony.Golia@MorganStanley.com
2005-Oct-25 15:09 UTC
Re: [Xen-users] veth0 is from netback and vifu.0 is fromthebridge-utils?
On Mon, 24 Oct 2005, Nivedita Singhvi wrote:> Anthony.Golia@MorganStanley.com wrote: > > > thx. that helps but leaves me with a few more questions. any doc is > > appreciated. i had bridging on then changed the xend-config.sxp > > to use vif-route so now i have a completely different set of ints > > (see below). i thought the eth0 in domU was linked to vifx.U in dom0? > > so with routing on there''s no peth0? is eth0 and eth1 still my real, > > physical interface? the ascii diagram you have above is helpfull how does > > it look with routing? > > > > bash-3.00# ip addr > > 1: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 > > link/ether 00:09:6b:b5:6b:94 brd ff:ff:ff:ff:ff:ff > > inet 172.31.205.177/24 brd 172.31.205.255 scope global eth0 > > 2: eth1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 1000 > > link/ether 00:09:6b:b5:6b:95 brd ff:ff:ff:ff:ff:ff > > inet 172.31.215.177/24 brd 172.31.215.255 scope global eth1 > > 3: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue > > 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 172.31.254.207/32 brd 172.31.254.207 scope global lo:hostname > > 4: vif0.0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop > > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > > 5: veth0: <BROADCAST,MULTICAST> mtu 1500 qdisc noop > > link/ether 00:00:00:00:00:00 brd ff:ff:ff:ff:ff:ff > > 6: vif1.1: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc noqueue > > link/ether fe:ff:ff:ff:ff:ff brd ff:ff:ff:ff:ff:ff > > inet 169.254.1.0/32 brd 169.254.1.0 scope global vif1.1 > > This really depends on your patchset/version you are running, and > your local script. Are you running the default network-route and > vif-route that''s currently in the unstable tree? If you change > the vif file to be vif-route, you should have changed the network > file to be network-route, too. Looks like you have two interfaces > in dom0, that''s why you have both eth0 and eth1. You have two > virtual backends (vif0.0, vif1.1). The default route scripts don''t > rename the interface (so you won''t have a peth0, for instance).. > Is the above working for you? Can you send netstat -rn output? > > In brief: > > physical interface (e.g. eth0) <=> virtual backend (e.g. vif0.0) > || > \\ > =====> domU nic (eth0) >thx. running changeset: 7421:aabc33c3c0ac with the default vif-route and network-route. i dont fully understand the tie between vif0.0 and eth0 in dom0. i thought vif1.1 in dom0 was linked to eth0 in domU (not vif0.0 in dom 0 as in ur diagram). yes i have two real ints eth0 and eth1 (and i intend to use both of them using linux''s advanced routing/load balancing). bash-3.00# ip route show 172.31.254.253 dev vif1.1 scope link < ***** i added this manually 172.31.215.128/25 dev eth1 proto kernel scope link src 172.31.215.177 172.31.205.128/25 dev eth0 proto kernel scope link src 172.31.205.177 172.31.205.0/24 dev eth0 proto kernel scope link src 172.31.205.177 172.31.215.0/24 dev eth1 proto kernel scope link src 172.31.215.177 169.254.0.0/16 dev eth1 scope link default nexthop via 172.31.205.1 dev eth0 weight 1 nexthop via 172.31.215.1 dev eth1 weight 1 despite my confusion (getting better thx to you) things seem to be working, mostly, cept some small thingies like: - xm create doesnt seem to like a 32 bit netmask (255.255.255.255) in the dom conf file - same for having dom1 be its own default router (i.e. proxy arp)> I''ll confirm this once I get the latest tip working... > > thanks, > Nivedita >Cheers, Anthony _______________________________________________ Xen-users mailing list Xen-users@lists.xensource.com http://lists.xensource.com/xen-users