[rabbitmq-discuss] Gracefully shutting down rabbitmq

Ben Hood 0x6e6562 at gmail.com
Mon Sep 1 21:33:40 BST 2008


On Mon, Sep 1, 2008 at 9:23 PM, Aman Gupta <rabbitmq at tmm1.net> wrote:
> Ah, very interesting. What changed that the recovery time is reduced so
> drastically?

Don't really know for certain, but we're refactoring the routing so
that it scales (in the worst case) linearly with the number of queues
and bindings, which in the current mainline it doesn't really do. The
re-arranging of the routing entries may have caused recovery to speed
up, but that certainly wasn't the intention, the intention was to
maintain some kind of sublinear complexity when routing to very large
numbers of queues and bindings under churn. If you want to check it
out, it's under bug18776 in hg, it passes all of the functional test
suites but we're still not convinced that it's quite there yet, hence
I'm still testing it. I'm hoping to land it this week some time, all
things going well.


More information about the rabbitmq-discuss mailing list