HA settings not allowing virtual machines to start

[JAIME SCHMIDTKE]

I have a xenserver pool that is responsible for serving up my xendesktop environment. I have enabled HA on the pool and protected all servers on my hosts. I have over 100 desktop VMs right now looking to add more in the near future. I have initially set the failure number to 2 hosts that the pool can withstand. I have noticed when the Desktop Delivery Controllers attempt to start more VMs in the xenserver pool that they are unable to start. When I try to manually start the VM I get the following error:

"This operation cannot be performed because HA would no longer be guaranteed for this pool. To perform this operation anyway, you must disable or reconfigure HA"

I lower my host failure number to 1 and I am able to start the VM successfully. This sounds like a bug to me because I have all my desktop VMs set to Do Not Restart. I have enough resources in the pool (memory) for 2 failures to take place and my protected servers to be running. Im worried that if I provision more desktops that the problem will continue and eventually force me to disable HA on my pool. Am I right in thinking that this shouldnt be happening?

im running 5.5 update 2

[Tobias Kreidl]

Jaime,
Assuming you really have enough resources such that the remaining servers in the pool not have enough capacity to accommodate all the "protected" fail-overs, then I'd imagine this type of error message could+ be generated if the server count is off, perhaps because XenCenter is confused, so the first thing I'd do is restart XenCenter. When you list two servers can fail, how many servers are then left and taking the remaining servers with the +smallest+ amount of memory, do you +still have enough capacity to be able to run all the "protected" VMs? You could also try to force the number of tolerated failures from 2 down to 1and see if that makes a difference.
--Tobias

[Paolo Geninazzi]

Hi,

Maybe I can help you...

In the MASTER open a shell and type:

xe pool-list

you will get the pool's list.

xe pool-param-list uuid=[uUID]

Where [uUID] is the "number" found before DON'T USE BRACKETS.

Here you receive the config, near the end look at the line HA-ALLOW-OVERCOMMIT, it should be on FALSE

Then run this:
xe pool-param-set ha-allow-overcommit=true uuid=[uUID]

and then again check:

xe pool-param-list uuid=[uUID]

Now you should be able to rejoin your host.

let me know

[Alexander Velikikh]

Quote

xe pool-param-set ha-allow-overcommit=true uuid=[uUID]

Is that command are restart any hosts in pool?