Hi,
I use a filesystem resource with an IP address to failover an NFS mount to
different nodes in the cluster. The NFS mount is use as a single source of
the executables, accessible from multiple App Servers. The one point of
failure was the NFS mount. By providing an FS+IP as a resource,
I'm sure others can provide a much more in depth answer but in short...
Not all "file systems" are cluster aware.
Something like GFS can be active on all nodes at the same time; because
it uses a lock manager that makes every node in your cluster "chat" with
each other to determine who is writ
Greetings,
On Tue, Mar 16, 2010 at 9:34 AM, Jonathan Horne wrote:
>
> what im not understanding is, whats the point of creating the
> file system as a service?
> what, is this also going to be passed back and forth between
> hosts?
Consider a customer who cannot afford Oracle RAC and chooses
act
ok, i can understand the usefulness of luci for creating clustered services.
yes, create the IP, add the init script of the daemon you want to pass back and
forth. i can build all that and it works fine.
what im not understanding is, whats the point of creating the file system as a
service? w
