Hi,
we set up a 2 node cluster using ocfs2 to build a load balanced cluster with
apache2. A customer added a high traffic project and now, when high traffic
peaks come in, the second node reboots. What i figured out so far by
researching through different forums and mailing lists, is that the problem
is ocfs2 fencing.
I wanted to increase the time till reboot, but i couldnt determine what
value should be changed and where? Since we're running on gentoo linux and
dont have an official build, most of the paths and filenames dont match. I
hope you can tell me where to put what file to change the time till fencing.
The second question is about the cluster.conf
At the moment its a 2 node cluster, but it will contain about 20 nodes
within the next few years. Therefore we formatted the iscsi harddisk for 20
nodes.
The cluster.conf looks like the followingL:
node:
ip_port = 7777
ip_address = 192.168.0.131
number = 0
name = XYZ1
cluster = ocfs2
node:
ip_port = 7777
ip_address = 192.168.0.132
number = 1
name = XYZ2
cluster = ocfs2
cluster:
node_count = 20
name = ocfs2
Does the node_count need to be 20 because we formatted the hd for 20 nodes
or do we have to set 2 here? Are there any drawbacks when using 20 instead
of 2?
Thanks for your time, I really appreciate it
Jürgen
_______________________________________________
Ocfs2-users mailing list
Ocfs2-users@oss.oracle.com
http://oss.oracle.com/mailman/listinfo/ocfs2-users
