Thanks for clarifying that. Does BackupManager support auto-scaling and
cycled restarts of all nodes (for web-app upgrades) without losing the
user's session? If the backup node is restarted...the backup is lost isn't
it?

On Fri, Apr 4, 2025 at 8:23 PM Mark Thomas <ma...@apache.org> wrote:

> On 04/04/2025 02:42, Chuck Caldarale wrote:
> >
> >> On 2025 Apr 3, at 19:57, Tim N <tnti...@gmail.com> wrote:
> >>
> >> For a long time up to the latest version 11 documentation, there has
> been a
> >> recommended maximum limit of 4 nodes per cluster.
> >>
> >> https://tomcat.apache.org/tomcat-11.0-doc/cluster-howto.html
> >> "This works great for smaller clusters, but we don't recommend it for
> >> larger clusters — more than 4 nodes or so."
> >>
> >> Are there any plans to improve this?
> >>
> >> It's a pity to have to change the architecture when going from, say, 3
> >> notes to 8 by introducing farms. What is the next simplest free cluster
> to
> >> move to? Redis? Any idea how cluster farming compares with redis?
> >>
> >> What other options are there?
> >
> >
> >
> > I may be misreading the documentation, but I think the 4-node
> restriction applies to the DeltaManager, and using the BackupManager
> removes the limitation.
>
> Chuck is correct. The issue with the DeltaManager is that the cluster
> traffic scales with the square of the number of nodes. For the
> BackupManager traffic scales linearly with the number of nodes.
>
> The limit of 4 is one of those values that should work with most
> applications. Depending on your application, the actual limit may be
> higher.
>
> Mark
>
>
>
> >
> >    - Chuck
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
> > For additional commands, e-mail: users-h...@tomcat.apache.org
> >
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org
> For additional commands, e-mail: users-h...@tomcat.apache.org
>
>

Reply via email to