[lxc-devel] NFS mounts inside a container uses/requires the host IPstack

Tim Spriggs tims at uahirise.org
Wed Mar 16 18:51:43 UTC 2011


Thanks for the offer but I think I have networking under control. What
is not working properly is that NFS happens from a host IP instead of
a context IP... even though it is started from the context IP.

On Wed, Mar 16, 2011 at 1:55 AM, Jäkel, Guido <G.Jaekel at dnb.de> wrote:
> Dear Tim,
>
> I guess you run into a problem with the default routes -- if you don't apply special things, there's only one default route! I struggled at start with the same, it think. But I solved it. And I don't need an ip assigned to the bridges.
>
> Please check at first, if  CONFIG_IP_ADVANCED_ROUTER  is enabled in your kernel.
>
> This give the possibility to define rule based routings, but you have to use the  ip  command (of iproute2) to set it up. You may google for such things. But I found, that on recent kernels (i'm using .37, too), I don't need to define such rules by myself. Instead of the kernel "does the right things(tm)" out of the box.
>
>
> That's my situation: I'm going to build up a completely real-diskless environment, a bladecenter with an external filer. Each blade is acting as a host for containers. And each container should be runnable on any host.
>
> * eth0 is reserved for the host blades to PXE boot and nfs-root from it. It's attached with an by the switch "enrolled" VLAN 163, i.e. it behave like ordinary network interface. This will make the bootstrapping much easier.
>
> * eth1 is used to supply the LXC. It's attached with 3 unenrolled VLAN's 160-102.
> * For each, an VLAN-Device named vlan16{0-2} is attached on eth1
> * For each, a bridge named br160{0-2} is started with the corresponding vlan16{0-2} if assigned.
> * In a container, a veth0 is linked to the desired bridge. It is set up (by DHCP) to use a IP on this VLAN and the default route for this VLAN. It also uses a virtual root filesystem. It is provided via LXC by the host, but as for the host, it points directly to a rootdir on the NFS.
>
> A traffic is routed in a correct manner. If I access to a subnet outside the Container's VLAN subnet, the switch will route it. And the communication partner sees the ip of the container.
>
>
> I'm using Gentoo Linux. If you need more details, please ask for.
>
> Greetings
>
> Guido
>
> --
> Dr. Guido Jäkel
> Deutsche Nationalbibliothek
> IT SG 2.2 (Infrastruktur Unix)
> Adickesallee 1
> 60322 Frankfurt am Main
> Tel. +49-69-1525-1750
> Fax  +49-69-1525-1799
> mailto:g.jaekel at dnb.de
> http://www.dnb.de
>
>
>
>
> ------------------------------------------------------------------------------
> Colocation vs. Managed Hosting
> A question and answer guide to determining the best fit
> for your organization - today and in the future.
> http://p.sf.net/sfu/internap-sfd2d
> _______________________________________________
> Lxc-devel mailing list
> Lxc-devel at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/lxc-devel
>
>




More information about the lxc-devel mailing list