|
On Mon, 9 Jul 2001, reddy karunakar wrote:
> I am Karunakar Reddy.p , doing my final semster of Master of Computer Science
> and Engineering.
> i am attending Anna University, Chennai,INDIA.
> I am intersted to work the following problem.
> The problem is:
>
> when primary load balancer fails, the back up server
> should continue the existing client requets.
> as i am doing academic project,please tell me whether it is possible
> to complete the work in four months.
I added a connection synchronization stuff into ipvs 0.9.2 (for kernel
2.4) a couple of weeks ago. It uses UDP multicast to propogate the
creation/changes of connections from the primary load balancer to a set of
backup load balancers. In order to keep the performance penalty of this
stuff small, the primary load balancer queues the changes of connections
in the sync queue, a syncmaster daemon started inside the kernel pipeline
the queue to the backups periodically. And, the syncbackup daemon is
started inside the kernel of the backup load balancers, it receives the
multicast message, and creates/changes ipvs connections respectively.
I don't know if there is a good alternative, maybe it is still worth
investigating on this issue.
Cheers,
Wensong
|
