On Feb 2, 2007, at 11:27 AM, Daniel Lopez wrote:

> Hi,
> Interesting. One question though, if I edit the config file and  
> modify one
> of the cluster blocks, to add a new app for example, is Resin smart  
> enough
> to just restart that specific cluster or will it restart all the  
> instances
> due to the new timestamp of the config file?
> I guess it is the latter, as
> detecting changes in the config file, checking if they affect one  
> or more
> instances etc... might be an overkill.

Resin would restart everything.  So that's an issue.  However, a  
resin:import should only affect the specific <cluster>, e.g.

     <resin:import path="..."/>

In that case, though, it might just be easier to have different  
resin.conf and therefore different watchdog-port values.

It might be possible to enhance the watchdog to keep the resin.conf  
information around, so you can use the same watchdog for different  
resin.conf as long as you keep the server identifiers unique.

-- Scott

> In that case, we would still need different config files as we  
> specify the
> context mappings explicitely, instead of using webapps default  
> mapping, and
> restarting 7-8 instances due to a change in one of the mappings  
> would not
> work for us. But I understand we are kind of a strange usecase,  
> with ~25
> different webapps on 6-7 instances with no clustering.
> Salute!
> D.
> Scott Ferguson wrote:
> ...
>>> And as Jose said, you don't need two copies of the whole Resin
>>> software,
>>> just two config files so you don't have to worry about
>>> maintaing/upgrading two full instances.
>> FYI, in Resin 3.1 you could just use one config file and have two
>> separate <cluster> blocks.  The -server foo will select which
>> <cluster> will be active for that JVM.
>> -- Scott
> _______________________________________________
> resin-interest mailing list
> resin-interest@caucho.com
> http://maillist.caucho.com/mailman/listinfo/resin-interest

resin-interest mailing list

Reply via email to