frank
|
2f634c0913
|
Switch to Apache license
|
2012-04-03 04:50:05 -07:00 |
Alex Huang
|
d01e20c443
|
bug 10094: The problem was we added code that won't add any more ha work items if it already has one. However, that is wrong. HA Manager stores the existing snapshot of the VM state machine. Before working on HA for a VM, it checks to see if that snapshot has been changed. So by not scheduling HA work, we've effectively made HA not work under multi-failure situations. I've fixed by removing that code and instead at the time of performing HA, do a quick check to see if there are pwork underway for the same VM and work scheduled in the future for that VM. If there are work scheduled in the future, then we simply cancel the current work. If there are already work underway, then we retry again in 1 minute.
|
2011-06-12 09:25:48 -07:00 |
|
Alex Huang
|
bee6953e09
|
ha done? needs testing
|
2011-02-07 10:58:25 -08:00 |
|
Alex Huang
|
d38f7fd56d
|
Moved DAO to server
|
2010-11-22 07:40:41 -08:00 |