<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
This is not a coherent shared disk environment. The iscsi target<br>
has no idea that the device is also being updated from another<br>
source.<br>
<br>
On 03/02/2011 04:30 PM, Nikola Savic wrote:
<blockquote cite="mid:4D6EE10A.7000001@logik-internet.rs"
type="cite">
<font size="-1"><font face="Helvetica, Arial, sans-serif"><br>
Hello,<br>
<br>
I have 3 node setup using CentOS 5.5 and OCFS2 1.4. Disks
from two
nodes (node1 and node2) are connected using DRBD
(primary/primary mode)
and this shared storage should be accessable from all three
nodes.
Because of that, node1 has the iSCSI target setup on top the
/dev/drbd1
device. I was also able to format the shared storage using
OCFS2.<br>
<br>
When I mount OCFS2 disk on node1 and node2 using the
/dev/drbd1
device, they work well. After mounting it on node3 using the
imported
iSCSI /dev/sda device, system gets into trouble. It looks like
node3
can't join cluster and it's heathbeating in slot taken by
node1 or
node2.<br>
<br>
Similarly, when I import iSCSI target from node1 on node2,
then node2
and node3 work well, while mounting /dev/drbd1 on node1 makes
hearthbeating problem. When that happens, on node2 and node3 I
get
following line in /var/log/messages:<br>
<i> node3 kernel: ocfs2_dlm: Nodes in domain
("72F96BF71A3A41B6BD5540CE773A9E9D"): 2 3</i><br>
while node1 reports:<br>
<i> node1 kernel: ocfs2_dlm: Nodes in domain
("72F96BF71A3A41B6BD5540CE773A9E9D"): 1<br>
...<br>
node1 kernel:
(o2hb-72F96BF71A,2456,0):o2hb_do_disk_heartbeat:777
ERROR: Device "drbd1": another node is heartbeating in our
slot!</i><br>
<br>
I tried to setup iSCSI initiator on node1, so it can connect
to own
iSCSI target, but it doesn't work.<br>
<br>
Does anyone know why is this happening and how can I solve
the
problem? I think everything worked well when I used GNBD to
export DRBD
device to third node, but I would like to do it without RHCS.<br>
<br>
Cluster configuration file (/etc/ocfs2/cluster.conf):<br>
cluster:<br>
node_count = 3<br>
name = testcluster<br>
<br>
node:<br>
ip_port = 7777<br>
ip_address = 192.168.100.2<br>
number = 1<br>
name = node1<br>
cluster = testcluster<br>
<br>
node:<br>
ip_port = 7777<br>
ip_address = 192.168.100.3<br>
number = 2<br>
name = node2<br>
cluster = testcluster<br>
<br>
node:<br>
ip_port = 7777<br>
ip_address = 192.168.100.4<br>
number = 3<br>
name = node3<br>
cluster = testcluster<br>
<br>
Thanks,<br>
Nikola<br>
</font></font>
<pre wrap="">
<fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
Ocfs2-users mailing list
<a class="moz-txt-link-abbreviated" href="mailto:Ocfs2-users@oss.oracle.com">Ocfs2-users@oss.oracle.com</a>
<a class="moz-txt-link-freetext" href="http://oss.oracle.com/mailman/listinfo/ocfs2-users">http://oss.oracle.com/mailman/listinfo/ocfs2-users</a></pre>
</blockquote>
<br>
</body>
</html>