Proposed F19 Feature: High Availability Container Resources

[Andrew Beekhof]

On 29/01/2013, at 7:46 AM, David Vossel <dvossel at redhat.com> wrote:

> 
> 
> ----- Original Message -----
>> From: "Bill Nottingham" <notting at redhat.com>
>> To: devel at lists.fedoraproject.org
>> Sent: Monday, January 28, 2013 1:18:06 PM
>> Subject: Re: Proposed F19 Feature: High Availability Container Resources
>> 
>> Jaroslav Reznik (jreznik at redhat.com) said:
>>> = Features/ High Availability Container Resources =
>>> https://fedoraproject.org/wiki/Features/High_Availability_Container_Resources
>>> 
>>> Feature owner(s): David Vossel <dvossel at redhat.com>
>>> 
>>> The Container Resources feature allows the HA stack (Pacemaker +
>>> Corosync)
>>> residing on a host machine to extend management of resources into
>>> virtual
>>> guest instances (KVM/LXC).
>>> 
>>> == Detailed description ==
>>> This feature is in response to the growing desire for high
>>> availability
>>> functionality to be extended outside of the host into virtual guest
>>> instances.
>>> Pacemaker is currently capable of managing virtual guests, meaning
>>> Pacemaker
>>> can start/stop/monitor/migrate virtual guests anywhere in the
>>> cluster, but
>>> Pacemaker has no ability to manage the resources that live within
>>> the virtual
>>> guests. At the moment these virtual guests are very much a black
>>> box to
>>> Pacemaker.
>>> 
>>> The Container Resources feature changes this by giving Pacemaker
>>> the ability
>>> to reach into the virtual guests and manage resources in the exact
>>> same way
>>> resources are managed on the host nodes. Ultimately this gives the
>>> HA stack
>>> the ability to manage resources across all the nodes in cluster as
>>> well as any
>>> virtual guests that reside within those cluster nodes.
>> 
>> Does this require the management to live on the virtual host, or can
>> it be
>> done entirely remotely with the cluster management server residing
>> elsewhere
>> and talking to all the virtual guest instances directly?
> 
> Management can be done entirely remotely from any cluster node running the ha stack.  There are no location restrictions. We are not restricted to the remote instance being a virtual guest either.  It could be bare-metal for all we care.
> 
> Initially the cli management tools and documentation will focus on virtual guest use case where the management is performed on the virtual host machine.  This just means we are planning on making that a very easy use-case to configure.  The tools will be available work outside of this use-case though.  It will just take a little more knowledge from the user.
> 

Just to explicitly call this out, we will be supporting two use-cases:
- whitebox, where a remote agent is installed on the guest (or non-clustered machine)
- blackbox, where there is _nothing_ installed on the guest (or non-clustered machine)

For the blackbox case you are obviously limiting yourself to testing externally exposed APIs to determine status and not being able to start/stop the services directly.
We are adding support for nagios scripts which seem popular for this task.

-- Andrew