<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=ISO-8859-1">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<font face="Calibri">Hi,<br>
<br>
I'm setting up a GlusterFS replicated cluster with CTDB and
build-in glusterfs nfs export on CentOS 6.4 hosts following the
documentation.<br>
Everything works fine, but I have a question about the setup:<br>
<br>
As there is currently no event script in ctdb for monitoring the
gluster status, a node will not become unhealthy if for some
reason the glusterd nfs export is unavailable.<br>
As such, the public ip stays at a node even when it's not serving
the gluster volume or when it's not in sync.<br>
<br>
Does anybody already had a go at an ctdb event script for
monitoring the gluster vol status on the host?<br>
<br>
regards,<br>
<br>
Sam</font>
</body>
</html>