hosted-engine and GlusterFS on Vlan help
Hi Guys, I've converted my lab from using 802.3ad with bonding>bridged vlans to one link with two vlan bridges and am now having traffic jumping to the gateway when moving VM's/ISO/etc. 802.3ad = node1>switch1>node2 801.1q = node1>switch1>gateway>switch1>node2 I assume I've setup the same vlan style, though this time I used the gui on the initial host install... setting up the vlans with their parent being eth0. Hosted-engine on deploy then creates ovirtmgmt on top of eth0.11 ... Switch is tagged for vlans 10 & 11. Including a PVID of 11 for good measure. (Gluster is vlan 11) I'd expect the traffic from node to node to be going from port to port like it did in 802.3ad, what have I done wrong or is it using the gui initially? This is how the current setup looks: /var/lib/vdsm/Persistent/netconf/nets/ovirtmgmt: { "ipv6autoconf": false, "nameservers": [], "nic": "eth0", "vlan": 11, "ipaddr": "10.0.3.11", "switch": "legacy", "mtu": 1500, "netmask": "255.255.255.0", "dhcpv6": false, "stp": false, "bridged": true, "gateway": "10.0.3.1", "defaultRoute": true } /etc/sysconfig/network-scripts/ifcfg-ovirtmgmt: # Generated by VDSM version 4.18.13-1.el7.centos DEVICE=ovirtmgmt TYPE=Bridge DELAY=0 STP=off ONBOOT=yes IPADDR=10.0.3.11 NETMASK=255.255.255.0 GATEWAY=10.0.3.1 BOOTPROTO=none DEFROUTE=yes NM_CONTROLLED=no IPV6INIT=no VLAN_ID=11 MTU=1500 Thanks!!
Running iperf3 between node1 & node2, I can achieve almost 10gbps without ever going out to the gateway... So switching between port to port on the switch is working properly on the vlan. This must be a problem in the gluster settings? Where do I start troubleshooting here? On 10/04/2016 10:38 AM, Hanson wrote:
Hi Guys,
I've converted my lab from using 802.3ad with bonding>bridged vlans to one link with two vlan bridges and am now having traffic jumping to the gateway when moving VM's/ISO/etc.
802.3ad = node1>switch1>node2 801.1q = node1>switch1>gateway>switch1>node2
I assume I've setup the same vlan style, though this time I used the gui on the initial host install... setting up the vlans with their parent being eth0.
Hosted-engine on deploy then creates ovirtmgmt on top of eth0.11 ...
Switch is tagged for vlans 10 & 11. Including a PVID of 11 for good measure. (Gluster is vlan 11)
I'd expect the traffic from node to node to be going from port to port like it did in 802.3ad, what have I done wrong or is it using the gui initially?
This is how the current setup looks:
/var/lib/vdsm/Persistent/netconf/nets/ovirtmgmt: { "ipv6autoconf": false, "nameservers": [], "nic": "eth0", "vlan": 11, "ipaddr": "10.0.3.11", "switch": "legacy", "mtu": 1500, "netmask": "255.255.255.0", "dhcpv6": false, "stp": false, "bridged": true, "gateway": "10.0.3.1", "defaultRoute": true }
/etc/sysconfig/network-scripts/ifcfg-ovirtmgmt: # Generated by VDSM version 4.18.13-1.el7.centos DEVICE=ovirtmgmt TYPE=Bridge DELAY=0 STP=off ONBOOT=yes IPADDR=10.0.3.11 NETMASK=255.255.255.0 GATEWAY=10.0.3.1 BOOTPROTO=none DEFROUTE=yes NM_CONTROLLED=no IPV6INIT=no VLAN_ID=11 MTU=1500
Thanks!!
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
On Tue, Oct 4, 2016 at 9:51 PM, Hanson <hanson@andrewswireless.net> wrote:
Running iperf3 between node1 & node2, I can achieve almost 10gbps without ever going out to the gateway...
So switching between port to port on the switch is working properly on the vlan.
This must be a problem in the gluster settings? Where do I start troubleshooting here?
Is it the gluster traffic that's routed via gateway? By moving VM, did you mean migrating VM from one host to another? In this case, there's no storage migration. Why do you think this is a problem in gluster settings?
On 10/04/2016 10:38 AM, Hanson wrote:
Hi Guys,
I've converted my lab from using 802.3ad with bonding>bridged vlans to one link with two vlan bridges and am now having traffic jumping to the gateway when moving VM's/ISO/etc.
802.3ad = node1>switch1>node2 801.1q = node1>switch1>gateway>switch1>node2
I assume I've setup the same vlan style, though this time I used the gui on the initial host install... setting up the vlans with their parent being eth0.
Hosted-engine on deploy then creates ovirtmgmt on top of eth0.11 ...
Switch is tagged for vlans 10 & 11. Including a PVID of 11 for good measure. (Gluster is vlan 11)
I'd expect the traffic from node to node to be going from port to port like it did in 802.3ad, what have I done wrong or is it using the gui initially?
This is how the current setup looks:
/var/lib/vdsm/Persistent/netconf/nets/ovirtmgmt: { "ipv6autoconf": false, "nameservers": [], "nic": "eth0", "vlan": 11, "ipaddr": "10.0.3.11", "switch": "legacy", "mtu": 1500, "netmask": "255.255.255.0", "dhcpv6": false, "stp": false, "bridged": true, "gateway": "10.0.3.1", "defaultRoute": true }
/etc/sysconfig/network-scripts/ifcfg-ovirtmgmt: # Generated by VDSM version 4.18.13-1.el7.centos DEVICE=ovirtmgmt TYPE=Bridge DELAY=0 STP=off ONBOOT=yes IPADDR=10.0.3.11 NETMASK=255.255.255.0 GATEWAY=10.0.3.1 BOOTPROTO=none DEFROUTE=yes NM_CONTROLLED=no IPV6INIT=no VLAN_ID=11 MTU=1500
Thanks!!
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
participants (2)
-
Hanson -
Sahina Bose