[Linux-cluster] GFS withdraw VIP and FS Quiesce

[Graeme Crawford]

Hi All,

A two part scenario/question:

1) We are running a 4 node GFS Cluster with RGmanager managed Virtual ip's(VIP):
What happens here is a node withdraws from the cluster and the
RGmanager VIPs dont fail over to other nodes as they are oblivious to
the gfs withdraw.

Is there anyway I can get the VIP's to relocate should a node withdraw.

2) Has anybody used the quiesce functionality with GFS we are looking
at using clones and as we fracture a clone we would like the data to
be consistent on disk right before we fracture. The data on disk is
file data, so no DB.

Thanks,

Graeme.