[ClusterLabs] GFS and cLVM fencing requirements with DLM

[Digimer]

On 15/03/16 02:12 PM, Ferenc Wágner wrote:
> Hi,
> 
> I'm referring here to an ancient LKML thread introducing DLM.  In
> http://article.gmane.org/gmane.linux.kernel/299788 David Teigland
> states:
> 
>     GFS requires that a failed node be fenced prior to gfs being told to
>     begin recovery for that node
> 
> which sounds very plausible as according to that thread DLM itself does
> not make sure fencing happens before DLM recovery, thus DLM locks could
> be granted to others before the failed node is fenced (if at all).
> 
> Now more than ten years passed and I wonder
> 
> 1. if the above is still true (or maybe my interpretation was wrong to
>    start with)

Yes, fencing is required. DLM will not begin recovery until it is told
that the lost node is confirmed fenced.

> 2. how it is arranged for in the GFS2 code (I failed to find it with
>    naive search phrases)
> 
> 3. whether clvmd does the same

Anything that uses DLM requires fencing, which includes clustered LVM.
(Really, all clusters require fencing to avoid split-brains).

> 4. what are the pros/cons of disabling DLM fencing (even the dlm_stonith
>    proxy) and leaving fencing fully to the resource manager (Pacemaker)

Pacemaker's fencing will inform DLM when the node has been terminated.
If EL6-based clusters, this is done via 'fence_pcmk' config in cman's
cluster.conf (which simply asks pacemaker to do the fence and report
back when successful).

-- 
Digimer
Papers and Projects: https://alteeve.ca/w/
What if the cure for cancer is trapped in the mind of a person without
access to education?