redo and update how we synchronize flags across the cluster.
authorroot <root@test1n1.VSOFS1.COM>
Fri, 5 Dec 2008 05:32:30 +0000 (16:32 +1100)
committerroot <root@test1n1.VSOFS1.COM>
Fri, 5 Dec 2008 05:32:30 +0000 (16:32 +1100)
commit32d460b8469eb53145f04161a5d01166f9b5f09e
treeff970abf2be0a6ab6a6fc227a0cbcb410c67ddfd
parent5d3018c37179966f75183d9a98790eaaaf1d2cfc
redo and update how we synchronize flags across the cluster.
this simplifies the code and should close a race condition between the local recovery daemon and a remote node when flags are changing.
server/ctdb_recoverd.c
tcp/tcp_connect.c
tools/ctdb.c