[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [Linux-cluster] fence device



Jonathan,

I know GNBD is a choice, but I will not consider it in my study environment.

Thank for you very much.

Michael

On 7/29/05, Jonathan E Brassow <jbrassow redhat com> wrote:
> 
> On Jul 28, 2005, at 9:18 PM, Q.L wrote:
> 
> > So, is there any 8 ports brocade switch that support fencing agent?
> 
> Pretty much all brocade switches are supported by the (included)
> fence_brocade agent.
> 
> > can Qlogic switch work?
> 
> I think so.  We have a fence_sanbox2 agent... not sure about all the
> models it supports - perhaps all, perhaps one.  Help, anyone?
> 
> >  If I use a switch without fencing agent, can I
> > use a manual fencing way to get system work?
> 
> Yes, but it will require manual intervention every time you have (or
> test) a failure.
> 
> >  In fact, what I want to
> > study is the Symmetric lock principle, especially at the time
> > concurrent write/read to the storage pool from both nodes, without
> > considering one point failure.
> 
> If you had a third machine, you could use GNBD.  The GNBD server would
> be a SPOF, but if it's just for testing, it would be fine.  GNBD would
> turn your third box into a iSCSI-like device which has built in I/O
> fencing support.
> 
>  brassow
> 
>


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]