From: Leandro Santi (lesantiuolsinectis.com.ar)
Date: Wed Jun 11 2003 - 16:33:29 CDT

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, Jun 11, 2003, Victor.Duchovnimorganstanley.com wrote:
> On Wed, 11 Jun 2003, Leandro Santi wrote:
>
> > On Wed, Jun 11, 2003, Wietse Venema wrote:
> > > Leandro Santi:
> > > > Oh. But... what about the proxymap? If someone uses the proxymap as a way
> > > > to save lots of MySQL connections, then the coin method can have somewhat
> > > > poor fairness performance (because the number of processes won't be
> > > > always high enough -statistically speaking, of course-).
> > >
> > > If you have only a few connections, who cares? Load balancing
> > > is an issue only when you have enough load to begin with.
> >
> > With the proxymap one could have relatively few instances (say, 5) of it,
> > so the number of connections will be low as well, but that does not
> > directly imply low load. Correct me if I'm wrong, but AFAIK multi server
> > processes like the proxymap don't enforce max_use strictly. This isn't good
> > for the coin solution.
> >
>
> Wietse's point is that lack of load-balancing is not important unless it
> leads to query latency. Query latency will cause the proxymap service to
> be busy and therefor more instances will be spawned. If 5 proxymap
> instances are all it takes to handle the load, the load-balancing is
> adequate.

Yes. I didn't think about this scenario, but now I see, and he's probably
right. Thanks.

Leandro.

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]