openstack team mailing list archive
-
openstack team
-
Mailing list archive
-
Message #06430
Re: Configure Rate limits on OS API
On Mon, Jan 2, 2012 at 6:27 AM, Soren Hansen <soren@xxxxxxxxxxx> wrote:
> 2011/12/29 Vishvananda Ishaya <vishvananda@xxxxxxxxx>:
>> In addition to the rate limiting changes, you can also run multiple copies
>> of nova-api on different ports and put a load balancer (like ha-proxy) in
>> front. You may have to set up your proxy to add an X-Forwarded-For header
>> and specify the
>> --use_forwarded_for
>> flag to get things like metadata to work.
>
> I can't think of any reason we can't just fork() N times (where N
> defaults to number of cores, but could be overridden by a flag) after
> socket.listen(), but before we go into the socket.accept() loop?
This is precisely what Swift does. And, very shortly, Glance:
https://review.openstack.org/#change,2486
Cheers,
-jay
Follow ups
References