[Pacemaker] Get group behaviour with Master slave or clones envolved

[Andrew Beekhof]

On 21 Jan 2014, at 10:50 pm, Néstor C. <xalabin at gmail.com> wrote:

> Hello.
> 
> When you need that some primitives switch in block you can use a group.

Groups are just a syntactic shortcut for ordering and colocation constraints.

> 
> There is a manner to get this when you have a clone or a master/slave involved?
> 
> For example:
> 
> Imagine a drbd disk (DR), a filesystem over ti (FS) and a service over all (SRV).
> The first one is a ms resource, and the other, primitives.
> 
> The colocation rules are:
> colocation fs_on_dr inf: FS DR:Master
> colocation srv_on_fs inf: SRV FS
> 
> The order rules are:
> order fs_after_dr inf: DR:promote FS:start
> order srv_after_fs inf: FS:start SRV:start
> 
> How can switch the entire cluster to other node if SRV fails? (like if all was in a group)

In what way does it not do so already with the above constraints?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 841 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: <http://lists.clusterlabs.org/pipermail/pacemaker/attachments/20140217/bf7da244/attachment-0002.sig>