Remove a host from a pool

mauzilla

Is it possible to remove a host from a pool and set it back to be managed by it's own?

olivierlambert

Sure. In XO:

Host view
Advanced tab
Detach

Double check you don't have any VM running on a local SR for this host before doing this operation.

tjkreidl

And from the CLI:

xe host-list (to get the UUID of the host)
xe pool-eject host-uuid=<host_UUID>

mauzilla

@olivierlambert I noted now when removing the host that is says it removes all VM's from the local storage and reboots. When you add the host to a pool, you can simply shutdown all VM's on the host, why does it remove vm's if you want to remove the host from a pool?

olivierlambert

Because it will create a fresh XAPI database. It doesn't know how to "extract" records that are specific to a host.

dalat9181

How long does it takes for the host to be remove from the pool?

olivierlambert

Something like between 10 and 30 seconds in average.

dalat9181

@olivierlambert
I migrate all the vm of the host server.
put it in maintained mode
and detach it. But it have spinning for about an hour or now.
The reason I want to detach it so I can re-install xcp-ng with ext storage instead of LVM.

olivierlambert

There's no need to reinstall to do this

Just remove the LVM SR, recreate it in ext and that's it.

dalat9181

@olivierlambert
Thanks Oliver!
Is there a document for this?

olivierlambert

To destroy the SR and recreate it?

dalat9181

@olivierlambert
Hi Oliver,
yes. I was just wondering if you have an official doc for it. But I did google on how to do this and will try it out.
Thanks

olivierlambert

In XO, it's pretty straightforward: Home/SR, delete the SR. Then, New/SR and recreate it

kevdog

@tjkreidl

So here is the dilemma I'm running into, the host I added with the incorrect network ports and such. It was reachable via the internet until I upgraded some pool patches, now its extremely unreachable. I can boot the installation, however I can't reach the slave to eject the host.

I tried emergency resetting network, however that didn't work either.

lshw -C network reveals a lot of unclaimed ethernet controllers which are the Intel 225-V version. I looked up various internet resources to see what "unclaimed" meant, and these sources suggest there isn't an appropriate driver -- which is weird since prior to the pool patches applied this installation was up and running (but misconfigured).

Sooo -- I either have two options at this point, and I don't really care. I'd like to eject this pool member -- destroy it -- however it seems I can't since it's unreachable, or I need to reinstall the I225V network driver.

Any suggestions?

tjkreidl

@kevdog You can force a host to be ejected even if the host is not reachable any longer from the pool master using:
xe host-forget uuid=UUID
If there is any issue because VMs are thought to be running still on that host,
you may need to do a power state reset on such before you can get rid of that host.
See if that works for you.