[Users] Logical networks in 3.1 on CentOS 6.3
Hi list, I'm testing oVirt on 3 hosts as a possible replacement for Proxmox. I've some trouble to understand the network part of the config (for beginning, I'll have some other questions about glusterfs later :)) and need some lights. The environment : - Engine host, no-virtualisation capability, one NIC eth0 - 2 nodes, virtualisation OK, 2 NICs eth0,eth1 - oVirt 3.1 on CentOS 6.3. The installation of engine/nodes was OK, no problem at all on the 3 nodes. The NIC eth0 on all hosts was manually bound on a bridge "ovirtmgmt", respectively on 10.0.0.1, .2, .3 ovirmgmt appears on the console as the management network, no problem, just a first question : it is marked as "VM network", and no way to change that. What if I want that this network to be dedicated on physical hosts only ? Second, I tried to create a logical network, call it "vmnetwork", linked it to datacenter and cluster. In the network interface configuration window on the hosts, I can link vmnetwork on eth0, but when validating an error occurs saying that eth0 is already linked to ovirmgmt. So I can't have management network and vm network on the same NIC , even on a virtual nic (tried with vlan=0 and vlan=10 on vmnetwork, same thing) ? Third, is this a bug when I come back to the configuration window, that the networks (for ovirmgmt and vmnetwork) return to "DHCP" ? If I forgot to re-check the static network, the host lose all it's network config and is declared non-responsive then rebooted..... Bad ! Thank you for helping me to understand. Frank
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Il 28/09/2012 18:43, Frank Soyer ha scritto:
[...] Third, is this a bug when I come back to the configuration window, that the networks (for ovirmgmt and vmnetwork) return to "DHCP" ? [...]
This behaviour apply exactly also on my side: every time I change the network config (adding/assigning VLANs to phisical NIC and/or to virtual networks - by-the-way of the "Setup Host Networks" form for each of my phisical hosts), the ovirtmgmt NIC config auto-reset to DHCP . If I forget to re-set it to the previously defined static IP address (by changing the radio-button in the "Edit Management Network" form), I loose connectivity to the phisical node (as I don't use DHCP on my management network). I wonder if something is wrong on my side, or if it's a bug to be filed somewhere. Bye, DV P.S. I'm also on a CentOS 6.3 platform. - -- Damiano Verzulli e-mail: damiano@verzulli.it - --- possible?ok:while(!possible){open_mindedness++} - --- "Technical people tend to fall into two categories: Specialists and Generalists. The Specialist learns more and more about a narrower and narrower field, until he eventually, in the limit, knows everything about nothing. The Generalist learns less and less about a wider and wider field, until eventually he knows nothing about everything." - William Stucke - AfrISPA http://elists.isoc.org/mailman/private/pubsoft/2007-December/001935.html -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAlBmwfwACgkQcwT9fsMT4SxsLQCgm0x38xspqGpxRiP7Tg2s+xpv 36YAoIwpb6R5d0b4uWdAl/llAmP5wx8K =u/qe -----END PGP SIGNATURE-----
On 09/28/2012 06:43 PM, Frank Soyer wrote:
Hi list, I'm testing oVirt on 3 hosts as a possible replacement for Proxmox. I've some trouble to understand the network part of the config (for beginning, I'll have some other questions about glusterfs later :)) and need some lights.
The environment : - Engine host, no-virtualisation capability, one NIC eth0 - 2 nodes, virtualisation OK, 2 NICs eth0,eth1 - oVirt 3.1 on CentOS 6.3.
The installation of engine/nodes was OK, no problem at all on the 3 nodes. The NIC eth0 on all hosts was manually bound on a bridge "ovirtmgmt", respectively on 10.0.0.1, .2, .3
ovirmgmt appears on the console as the management network, no problem, just a first question : it is marked as "VM network", and no way to change that. What if I want that this network to be dedicated on physical hosts only ?
On 3.2 Data Center you should be able to change the VM Network property of the ovirtmgmt network as long there are no clusters defined on the DC. In order to modify network property on DC level it should be detached from the DC's clusters, however since the management network cannot be detached from a cluster you should configure it prior to adding clusters to the DC. Update non-management network could be done by detaching the network from all clusters, modifying the network and re-assign the network to the clusters.
Second, I tried to create a logical network, call it "vmnetwork", linked it to datacenter and cluster. In the network interface configuration window on the hosts, I can link vmnetwork on eth0, but when validating an error occurs saying that eth0 is already linked to ovirmgmt. So I can't have management network and vm network on the same NIC , even on a virtual nic (tried with vlan=0 and vlan=10 on vmnetwork, same thing) ?
You could do that by defining the management network as vlan (requires configuring the network manually on the host) and set the vlan-tag on the ovirtmgmt network on DC's logical network sub-tab. Alternately, you should be able to attach networks with vlan-tag to an interface which a non-vm network (bridgeless) is already attached to it.
Third, is this a bug when I come back to the configuration window, that the networks (for ovirmgmt and vmnetwork) return to "DHCP" ? If I forgot to re-check the static network, the host lose all it's network config and is declared non-responsive then rebooted..... Bad !
What cluster level does the host member of? Is this behaviour in 'Setup Networks' or on 3.0 cluster?
Thank you for helping me to understand. Frank _______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
Hello, Le 30/09/2012 00:49, Moti Asayag a écrit :
Third, is this a bug when I come back to the configuration window, that the networks (for ovirmgmt and vmnetwork) return to "DHCP" ? If I forgot to re-check the static network, the host lose all it's network config and is declared non-responsive then rebooted..... Bad !
What cluster level does the host member of? Is this behaviour in 'Setup Networks' or on 3.0 cluster?
This is a v3.1 of oVirt (engine + nodes, on Centos 6.3), in the GUI when modifying the network interfaces of a node.
Hi, The bug (ip protocol is not saved) was fixed in previous commits. Use the latest oVirt version to get all the bug fixes. Attached a screen-cast that demonstrates how it works- http://www.youtube.com/watch?v=3EMrqQR-f3w Alona. ----- Original Message -----
From: "Frank Soyer" <fsoyer@systea.net> To: "Moti Asayag" <masayag@redhat.com> Cc: users@ovirt.org Sent: Tuesday, October 2, 2012 10:14:47 AM Subject: Re: [Users] Logical networks in 3.1 on CentOS 6.3
Hello,
Le 30/09/2012 00:49, Moti Asayag a écrit :
Third, is this a bug when I come back to the configuration window, that the networks (for ovirmgmt and vmnetwork) return to "DHCP" ? If I forgot to re-check the static network, the host lose all it's network config and is declared non-responsive then rebooted..... Bad !
What cluster level does the host member of? Is this behaviour in 'Setup Networks' or on 3.0 cluster?
This is a v3.1 of oVirt (engine + nodes, on Centos 6.3), in the GUI when modifying the network interfaces of a node. _______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
participants (4)
-
Alona Kaplan -
Damiano Verzulli -
Frank Soyer -
Moti Asayag