Hi Zen Masters,
We've reproduced in our lab a weird clustering behaviour: with "both can be
masters" cluster type, a backup node with a lower IP address (ex.
172.16.1.100) could be promoted itself to master, and the master with a
bigger IP address (ex. 172.16.1.101) demote itself, without any networking
issue.
*Virtual machines without the properly performance tools could cause a
clock drift which it's very negative for clustering synchronization*. ZenLB
executes NTP synchronization every hour, if the clock offset is bigger than
3 seconds the CARP service it's going to prepare the switch, as it suppose
that the master has been down for more than this value. During the
negotiation, two masters are found in the network, so the CARP algorithm
select the lower IP address to be the real master demoting the bigger IP
address node.
I've found a good document about timekeeping in Vmware
http://www.vmware.com/files/pdf/Timekeeping-In-VirtualMachines.pdf
Regards,
Laura.
------------------------------------------------------------------------------
Keep Your Developer Skills Current with LearnDevNow!
The most comprehensive online learning library for Microsoft developers
is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
Metro Style Apps, more. Free future releases when you subscribe now!
http://p.sf.net/sfu/learndevnow-d2d
_______________________________________________
Zenloadbalancer-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/zenloadbalancer-support
