On 6/26/06, James Pan <[EMAIL PROTECTED]> wrote:
I configured two resource groups namely "nfs" and "fs",
group "fs" contains one primitive resource: resource_fs which is a
filesystem,
the filesystem's device is set to an invalid device so that this
resource will always fail to start.
grou "nfs" contains one primitive resource resource_nfs which is a nfs
server.
this is the result of crm_resource -L:
-----------------------------------------------------------------
hadev1:~ # crm_resource -L
Resource Group: nfs
resource_nfs (lsb:nfsserver)
Resource Group: fs
resource_fs (heartbeat::ocf:Filesystem)
------------------------------------------------------------------
As the result, resource_nfs will be running and resource_fs will be stopped.
Now I want these two resources to be colocated, so i add a colocation
constraint for them.
if I add the following colocation constraint:
<rsc_colocation id="colocation_nfs_fs" from="nfs" to="fs" score="INFINITY"/>
nothing will change, resource_nfs will be still running and resources_fs
will be still stopped.
but if the constraint is:
<rsc_colocation id="colocation_nfs_fs" from="resource_nfs"
to="resource_fs" score="INFINITY"/>
this time resource_nfs will be stopped.
The only difference between the two constraints is the resource ids. one
uses the group ids and the other uses
the primitive resource ids. But the running result is different.
Is it a bug? It is easy to reproduce.
both work for me.
can you provide a configuration that does not?
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
