Hi All,
Anyone have experience in configuring 2 service consoles on different subnets? Specifically for the use of HA and its heartbeat?
Following the advice from the VMware HA Concepts, Implementation and Best Practices guide (Pg 23/24) it recommends that I configure a Secondary Service Console Network. I have done this on the vSwitch that also is home to the VMKernel Port (see attachment MyNetwork.JPG).
Primary ServiceConsole IP: 192.168.27.78
Primary ServiceConsole GW: 192.168.27.1
Secondary ServiceConsole IP: 192.168.26.178
Secondary ServiceConsole GW: 192.168.26.1
Advance HA options: das.failredetecttime 20000 and das.isolationaddress2 192.168.26.1
The properties of "Service Console 2" state that the default gateway is 192.168.27.1 on vswif0. However shouldn't this be 192.168.26.1 on vswif1?
It is reporting this information based on the configuration of the "Service Console 1".
I can only assume that routing for HA would be impossible if the primary Service Console nic/switch died as the Secondary Service Console would be trying to route through the primary 192.168.27.1 gateway, not its own on 192.168.26.1? After all, they are on seperate VLANs.
How do I configure the gateway correctly for the "Service Console 2" to allow for HA heart beat to still be active in light of a primary device failure.
I appreciate any help in advance.
Michael
You don't need to set a gateway for the second service console because you configured the das.isolationaddress2 address. when the first SC collapses the HA heartbeat will check if it's still alive. This will be a heartbeat to the das.isolationaddress2 address, and will be directly via the second service console cause that has the same subnet.
For RParker: http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=1002641&sl...
But just test it with 1 server with 1 test vm, just pull the primary SC network plug and see what happens, and put it back after 5 minutes and see what happens. Should just work fine with this setup.
Duncan
My virtualisation blog:
Not possible. ESX isn't designed for SC failover.
The secondary SC is only for SSH, should the primary go offline, so it allows you to "backdoor" the ESX host, and reboot if need be, but HA/DRS will not fail over to the secondary SC.
192.168.27.1 is correct. You can only default to the gateway on the SAME subnet.
You don't need to set a gateway for the second service console because you configured the das.isolationaddress2 address. when the first SC collapses the HA heartbeat will check if it's still alive. This will be a heartbeat to the das.isolationaddress2 address, and will be directly via the second service console cause that has the same subnet.
For RParker: http://kb.vmware.com/selfservice/microsites/search.do?cmd=displayKC&docType=kc&externalId=1002641&sl...
But just test it with 1 server with 1 test vm, just pull the primary SC network plug and see what happens, and put it back after 5 minutes and see what happens. Should just work fine with this setup.
Duncan
My virtualisation blog:
Thanks for explaining this Duncan.
Sorry about posting on an already answered thread..... but struggling to get this question answered.
djdeceit - Did your tests work? I have a similar confusing issue.... the same as you mentioned.... my thoughts would have been the second isolation address would not be reachable as to get to the second isolation address it has to route through the primary service consoles gateway (and this is not available as the vlan is shutdown for example).
Is that correct/incorrect Duncan? Or as you mentioned.... since the secondary isolation address is on the same subnet as the secondary SC..... it does not need to route via the primary SC gateway?
Well I got a chance to test this today.... HA didn't kick in so it was all good.