richard at catalyst
Aug 17, 2009, 3:31 PM
Post #3 of 3
(743 views)
Permalink
|
On Mon, 2009-08-17 at 18:50 +0900, Motoharu Kubo wrote:
> Hi Richard,
>
> > What is the address specified in the stacked-on-top-of section?
>
> The good reading is the DRBD Users Guide at www.drbd.org.
>
> There is a figure about concept of three-way replication in PartI
> "Introduction to DRBD"
> (http://www.drbd.org/users-guide-emb/s-three-way-repl.html).
>
> Figure 2.1 "DRBD resource stacking" shows that "stacked-on-top-of" side
> (the upper layer) DRBD must runs on top of PRIMARY lower layer DRBD.
> Therefore the IP address of "stacked-on-top-of" section should be one of
> floating address (virtual address) controlled by cluster manager.
>
> Please also refer to section "Creating a three-node setup" in Part III
> (http://www.drbd.org/users-guide-emb/s-three-nodes.html).
>
> I usually use eth1 for lower layer DRBD and eth0 for upper layer DRBD as
> well as service. For example, when using example configuration in Part
> III, 10.0.0.0/24 belongs to eth1. The eth0 address of alice and bob
> would be 192.168.42.100 and 192.168.42.101, respectively. My cluster
> manager should assign 192.168.42.1 to eth0:0 of active node.
Right, thanks. I'd seen all those, but obviously not looked at the
diagram carefully enough, where it shows the stacked drbd pointing at
the primary.
I don't intend to use any cluster manager, since I don't want any kind
of automated failover (or is that not a logical step?), so I'll
configure the address of the primary.
Assuming I am using the same interface for both (which I will
temporarily), presumably I'll need a different port number for the
stacked resource?
Thanks,
Richard
|
signature.asc
