[Linux-cluster] Split-brain with DRBD active-active + RHCS

jayesh.shinde jayesh.shinde at netcore.co.in
Tue Mar 15 12:13:14 UTC 2011

Hi All ,

I don't have SAN with me , so I want to build the 2 node DRBD active 
active for mysql & http  resource ( i.e /dev/drbd2 & /dev/drbd3 in my 
case) with RHCS .

I configured the require setup from 
http://sourceware.org/cluster/wiki/DRBD_Cookbook and from DRDB links.

 From last 1 week I am testing the same scenario in 2 XEN vms with 
kenel  2.6.18-128.el5xen , Every thing is working fine , like mysql and 
http services move from one server to other etc... But not working 
correctly when it get fence ( i.e when n/w fail on of the node).

*I am facing the split-brain problem. * I search a lot in google and 
mailling list but don't found the proper correct solution and suggestion.

For Fence testing  I am doing following.
1) On node1 http service is running with /dev/drbd2
2) On node2 mysql service is runing with /dev/drbd3
     At this movement the DRBD primary-primary status is working correctly.
3) Now On node2 When I stop n/w service manually by "service network 
stop"  then within 3-5 sec.  node2 get fence properly and mysql service 
get switch on node1 properly.
4) After fencing when node2 come up , then I am facing the DRBD 
split-brain issue with node1 and node2.

My questions :--
1) Why DRBD Split brain is not coming when I reboot or shutdown or 
destroy the machine by xm command
     i.e  "xm reboot <node1/node2>"   OR         xm shutdown 
<node1/node2>  OR      xm destroy <node/node2>

2) Why the DRBD split brain issue come at the time of fencing node only ?

3) Is the combination of DRBD active-active + RHCS is stable ? and 
workable solution
     Because one of the below mailling list I found it's workable solution

4) In fencing Is there any extra setting require for such combination  ?
5) Do I need to use any custom fencing logic ?

6) Any one using such "*DRBD active-active + RHCS*"  setup in Live 
without split brain issue ?

Please guide and suggest on the same.

Jayesh Shinde

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20110315/58cccc6d/attachment.htm>

More information about the Linux-cluster mailing list