On Thu, Mar 24, 2011 at 11:52 AM, Martin Pool <email address hidden> wrote:
> Of course this is a different xmlrpc (the internal dc one) to the
> other aspect of the bug.
Internally we have little separation: the internal xmlrpc is just
another port on our primary appservers. There is a different haproxy
cluster so the queueing effects can be different, but its suboptimal -
we'd like to move to a single optimisable config but thats a bit
longer term.
On Thu, Mar 24, 2011 at 11:52 AM, Martin Pool <email address hidden> wrote:
> Of course this is a different xmlrpc (the internal dc one) to the
> other aspect of the bug.
Internally we have little separation: the internal xmlrpc is just
another port on our primary appservers. There is a different haproxy
cluster so the queueing effects can be different, but its suboptimal -
we'd like to move to a single optimisable config but thats a bit
longer term.