Hello everyone
I had some strange problems on a ESX with two service console ports configured. I set up the host with a main service console at the ip xxx.xxx.133.1, then I added another service console with the ip xxx.xxx.133.2 to the same vswif (vswif0). Here I have 4 more server that are configured the same way with two SC in the same IP range. Suddenly I had the situation that i couldn't longer open a VM Console in Virtual Center or VI Client. Then if I changed the vswif gateway device like auto to vswif0 or vswif1 to vswif2 it worked again. Sunddenly it was broken again. Then I could go back to the vswif gateway and tinker again, then it will work again.
Now i logged in the second interface, deleted the first interface and made a new one. Now it works fine, but why. I can't understand, why this happend!
Here are some facts about the system:
root@xyz root]# esxcfg-vswif -l
Name Port Group IP Address Netmask Broadcast Enabled DHCP
vswif0 Service Console xxx.xxx.133.1 255.255.255.0 xxx.xxx.133.255 true false
vswif1 Service Console 2 xxx.xxx.133.2 255.255.255.0 xxx.xxx.133.255 true false
Switch Name Num Ports Used Ports Configured Ports Uplinks
vSwitch0 32 6 32 vmnic1,vmnic0
PortGroup Name Internal ID VLAN ID Used Ports Uplinks
Management VLAN 2 portgroup1 2 0 vmnic0,vmnic1
Service Console 2 portgroup12 2 1 vmnic0,vmnic1
Service Console portgroup8 2 1 vmnic0,vmnic1
VMkernel portgroup3 2 1 vmnic0,vmnic1
Switch Name Num Ports Used Ports Configured Ports Uplinks
vSwitch1 64 7 64 vmnic5,vmnic4,vmnic3,vmnic2
PortGroup Name Internal ID VLAN ID Used Ports Uplinks
Server Netzwerk portgroup4 8 2 vmnic2,vmnic3,vmnic4,vmnic5
I think that it is possible, that the was a bug in the frist service console, that made the second console take over. But which bug? All the other systems worked well. The Guest OS Console works on that interface which is in the VC, the other service console does work for all tasks, except this one.
How did you solve the second service console problem?
Do you have a second console or will you make one, when the first one is no longer available?
Thanks for all answers!
Greetings DLS
I wouldn't rely on this kind of redundancy when you setup a second SC on the same vswitch with the same subnet...?
If the first SC fails the second will not take over if it's in the same subnet, it's a known bug which causes the arp to be resolved to the same nic. I would setup a second service console port on the VMotion portgroup with the same subnet as the vmotion portgroup...
http://www.yellow-bricks.com/2008/01/14/service-console-redundancy/
Duncan
My virtualisation blog:
Can I ask why you have created 2 Service console port groups on the same vswitch?
The only time I can think of you needing a 2nd service port group is when you are using iscsi and even then, it is best practice to use a seperate vswitch on a dedicated LAN.
I did that because I want to bring in redudance for the Service Console. It used to be very useful when I had to change the IP of the main Service Console.
Yes, i know, there is a much easier way trought the CLI, but the Customer wishes it like that.
Documentation is like sex: when it is good, it is very, very good; and when it is bad, it is better than nothing.
I wouldn't rely on this kind of redundancy when you setup a second SC on the same vswitch with the same subnet...?
If the first SC fails the second will not take over if it's in the same subnet, it's a known bug which causes the arp to be resolved to the same nic. I would setup a second service console port on the VMotion portgroup with the same subnet as the vmotion portgroup...
http://www.yellow-bricks.com/2008/01/14/service-console-redundancy/
Duncan
My virtualisation blog:
If you are going to set up a secondary COS, it has to be an another vswitch for proper redundancy. Also, there has to be common gateway. Attached is a configuration we had setup. If the main Service Console is isolated for whatever reason, the secondary is just there so that the rest of Cluster doesn't know a host went down, thus causing an HA failure. The secondary service console will allow all the VM's to stay on-line even if the primary COS fails.
Why does it have to be a second vswitch. If I leave it on the same vswitch, i can use the second NIC for failover redundancy. I can still add another port group in a different subnet with another gateway device for HA. I have a stack with two cisco switches, each console NIC is connected to a switch. The VM networks are connected to the same stack, so if the stack goes down, it's not necessery that the HA stays online.
Is the das.isolationadress2 = xxx a parameter of 3.5 or is it also available in 3.0.2?
Documentation is like sex: when it is good, it is very, very good; and when it is bad, it is better than nothing.