[Sheepdog] [PATCH 0/3] fix master-failure logic

[Liu Yuan]

When the master node is down before new nodes finish joining, we have to 
revoke its mastership. Or the whole cluster get hanged.

Thanks
Yuan

Liu Yuan (3):
      cluster, corosync: add two dprintf
      cluster, corosync: enlarge is_master() audience
      cluster, corosync: do mastership transfer when master is down in join phase

 sheep/cluster/corosync.c |   34 ++++++++++++++++++++++++++++++----
 1 files changed, 30 insertions(+), 4 deletions(-)