From duelin.markers at gmail.com  Wed Jan  2 15:13:45 2008
From: duelin.markers at gmail.com (John D. Hume)
Date: Wed, 2 Jan 2008 15:13:45 -0500
Subject: [Rubyworks-users] capistrano recipes ... rubyworks_deployer?
Message-ID: <ce1ffa990801021213y6153da90w6eb8dd42cbe2b7d0@mail.gmail.com>

We're looking to set up a fresh Capistrano deployment for a single app
running on the RubyWorks stack (i.e., the simplest possible RubyWorks
deployment).  We've gone through the archives from this list looking
for links to recipes and haven't found any.  In fact all we've got is
the pretty reasonable tip to make /usr/rails a symlink to <capistrano
deploy_to>/current.
http://rubyforge.org/pipermail/rubyworks-users/2007-August/000047.html

Did we miss anything?  Anyway, for the time being we're piecing it
together ourselves.

Googling for "rubyworks capistrano" also turned up a rails plugin in
the RubyWorks svn trunk that seems to support remote installation of
the stack itself and presumably also deploying the Rails application
in which it's installed.
http://rubyworks.rubyforge.org/svn/trunk/rubyworks_deployer/
There hasn't been any discussion of that on the list.  What's the
story on this plugin?  Is it usable?

Thanks.
-hume.

From aslak.hellesoy at gmail.com  Wed Jan  2 18:15:53 2008
From: aslak.hellesoy at gmail.com (aslak hellesoy)
Date: Thu, 3 Jan 2008 00:15:53 +0100
Subject: [Rubyworks-users] why runit?
Message-ID: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>

Hi,

I'm digging into the RubyWorks stack and learning about monit and
runit. From what I understand, monit is capable of starting and
stopping processes and services, so I'm wondering:

Why is runit used to start services when monit can also do it?

Aslak

From aslak.hellesoy at gmail.com  Thu Jan  3 05:09:43 2008
From: aslak.hellesoy at gmail.com (aslak hellesoy)
Date: Thu, 3 Jan 2008 11:09:43 +0100
Subject: [Rubyworks-users] why runit?
In-Reply-To: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
Message-ID: <8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>

On Jan 3, 2008 12:15 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> Hi,
>
> I'm digging into the RubyWorks stack and learning about monit and
> runit. From what I understand, monit is capable of starting and
> stopping processes and services, so I'm wondering:
>
> Why is runit used to start services when monit can also do it?
>

Just to clarify: It seems to me that mongrels can be kept alive with:
  monit -> mongrel_cluster

Rubyworks is doing (as far as I can tell):
  monit -> runit -> mongrel_cluster

Why is this? I also came over a year old blog post that describes
monit and mongrel_cluster without any runit:
http://software.pmade.com/blogs/ramblings/2006/12/27/mongrel-cluster-and-monit

Aslak

> Aslak
>

From alexey.verkhovsky at gmail.com  Thu Jan  3 07:28:39 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Thu, 3 Jan 2008 05:28:39 -0700
Subject: [Rubyworks-users] why runit?
In-Reply-To: <3945c4270801030427y50254b43m41885e89e0945167@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<3945c4270801030427y50254b43m41885e89e0945167@mail.gmail.com>
Message-ID: <3945c4270801030428i7b382830s12c00bcd3f975cb6@mail.gmail.com>

On Jan 3, 2008 3:09 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> I'm digging into the RubyWorks stack and learning about monit and

Hey, Aslak,

> runit. From what I understand, monit is capable of starting and stopping processes and services

Monit is capable of executing commands to start and stop things.
Expectating those commands to create processes as proper *nix daemons.
Mongrel had a bunch of small but annoying problems with this aspect
(google for "mongrel pid" for a sample), and runit relatively neatly
gets around those problems. "Relatively neatly" because this solution
does create some duplication and complexity.

And no, mongrel_cluster is not part of the picture at all.
mongrel_cluster's function is performed by runit.

--
Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From gigix1980 at gmail.com  Thu Jan  3 07:33:38 2008
From: gigix1980 at gmail.com (Jeff Xiong)
Date: Thu, 3 Jan 2008 20:33:38 +0800
Subject: [Rubyworks-users] why runit?
In-Reply-To: <8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
Message-ID: <20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>

Monit starts mongrel and haproxy processes in daemon mode. That means
you have to manage all those PID files, otherwise you may get some
zombie PID files and cannot restart mongrel/haproxy any more. That's
why we added runit: it starts those processes as its sub-processes and
doesn't need any PID file.

So we just use monit to "monitor" mongrel/haproxy processes, and use
runit to "run it". It sounds reasonable, doesn't it?

On Jan 3, 2008 6:09 PM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> On Jan 3, 2008 12:15 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > Hi,
> >
> > I'm digging into the RubyWorks stack and learning about monit and
> > runit. From what I understand, monit is capable of starting and
> > stopping processes and services, so I'm wondering:
> >
> > Why is runit used to start services when monit can also do it?
> >
>
> Just to clarify: It seems to me that mongrels can be kept alive with:
>   monit -> mongrel_cluster
>
> Rubyworks is doing (as far as I can tell):
>   monit -> runit -> mongrel_cluster
>
> Why is this? I also came over a year old blog post that describes
> monit and mongrel_cluster without any runit:
> http://software.pmade.com/blogs/ramblings/2006/12/27/mongrel-cluster-and-monit
>
> Aslak
>
>
> > Aslak
> >
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users
>



-- 
Jeff Xiong
Software Journeyman - http://gigix.thoughtworkers.org
Open Source Contributor - http://stomperl.googlecode.com/
Technical Evangelist - http://www.infoq.com/cn/

From aslak.hellesoy at gmail.com  Thu Jan  3 08:25:17 2008
From: aslak.hellesoy at gmail.com (aslak hellesoy)
Date: Thu, 3 Jan 2008 14:25:17 +0100
Subject: [Rubyworks-users] why runit?
In-Reply-To: <20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
Message-ID: <8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>

On Jan 3, 2008 1:33 PM, Jeff Xiong <gigix1980 at gmail.com> wrote:
> Monit starts mongrel and haproxy processes in daemon mode. That means
> you have to manage all those PID files, otherwise you may get some
> zombie PID files and cannot restart mongrel/haproxy any more. That's
> why we added runit: it starts those processes as its sub-processes and
> doesn't need any PID file.
>
> So we just use monit to "monitor" mongrel/haproxy processes, and use
> runit to "run it". It sounds reasonable, doesn't it?
>

(Reposting - you should set the REPLY-TO header for this list to
rubyworks-users at rubyforge.org)

Jeff, Alex,

I thought it was the daemon/pid issue - just wasn't sure. Thanks for
clarifying this.

I've set up a clean ubuntu/rubyworks box at our intranet where we plan
to host a dozen or so Rails apps with very low traffic.
I'm thinking haproxy is not needed and one mongrel per app is probably enough.

I'm going to use monit/runit (and vlad) to get these apps running.

I'll probably make some vlad (Rake) plugins that will:

* Generate and upload a new runit config for a mongrel
* Generate and upload a new monit config for a mongrel
* Generate and upload a new Apache config that will set up proxying
via a virtual host
* Plus a standard way to redeploy applications and restart them with monit

Is this something that you'd be interested in including in the stack
if I donate it?

Aslak

>
> On Jan 3, 2008 6:09 PM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > On Jan 3, 2008 12:15 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > > Hi,
> > >
> > > I'm digging into the RubyWorks stack and learning about monit and
> > > runit. From what I understand, monit is capable of starting and
> > > stopping processes and services, so I'm wondering:
> > >
> > > Why is runit used to start services when monit can also do it?
> > >
> >
> > Just to clarify: It seems to me that mongrels can be kept alive with:
> >   monit -> mongrel_cluster
> >
> > Rubyworks is doing (as far as I can tell):
> >   monit -> runit -> mongrel_cluster
> >
> > Why is this? I also came over a year old blog post that describes
> > monit and mongrel_cluster without any runit:
> > http://software.pmade.com/blogs/ramblings/2006/12/27/mongrel-cluster-and-monit
> >
> > Aslak
> >
> >
> > > Aslak
> > >
>
> > _______________________________________________
> > Rubyworks-users mailing list
> > Rubyworks-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rubyworks-users
> >
>
>
>
> --
> Jeff Xiong
> Software Journeyman - http://gigix.thoughtworkers.org
> Open Source Contributor - http://stomperl.googlecode.com/
> Technical Evangelist - http://www.infoq.com/cn/
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users
>

From andy at adveho.net  Thu Jan  3 08:51:36 2008
From: andy at adveho.net (Andy Goundry)
Date: Thu, 3 Jan 2008 13:51:36 +0000
Subject: [Rubyworks-users] why runit?
In-Reply-To: <8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
Message-ID: <656150BB-2129-46F2-B438-4FC8261DDA0C@adveho.net>

Could Capistrano add anything to the mix here? I haven't used  
Rubyworks in anger yet, but capistrano is very customisable so could  
possibly handle your custom rollout and remote system management tasks.

Andy

On 3 Jan 2008, at 13:25, "aslak hellesoy" <aslak.hellesoy at gmail.com>  
wrote:

> On Jan 3, 2008 1:33 PM, Jeff Xiong <gigix1980 at gmail.com> wrote:
>> Monit starts mongrel and haproxy processes in daemon mode. That means
>> you have to manage all those PID files, otherwise you may get some
>> zombie PID files and cannot restart mongrel/haproxy any more. That's
>> why we added runit: it starts those processes as its sub-processes  
>> and
>> doesn't need any PID file.
>>
>> So we just use monit to "monitor" mongrel/haproxy processes, and use
>> runit to "run it". It sounds reasonable, doesn't it?
>>
>
> (Reposting - you should set the REPLY-TO header for this list to
> rubyworks-users at rubyforge.org)
>
> Jeff, Alex,
>
> I thought it was the daemon/pid issue - just wasn't sure. Thanks for
> clarifying this.
>
> I've set up a clean ubuntu/rubyworks box at our intranet where we plan
> to host a dozen or so Rails apps with very low traffic.
> I'm thinking haproxy is not needed and one mongrel per app is  
> probably enough.
>
> I'm going to use monit/runit (and vlad) to get these apps running.
>
> I'll probably make some vlad (Rake) plugins that will:
>
> * Generate and upload a new runit config for a mongrel
> * Generate and upload a new monit config for a mongrel
> * Generate and upload a new Apache config that will set up proxying
> via a virtual host
> * Plus a standard way to redeploy applications and restart them with  
> monit
>
> Is this something that you'd be interested in including in the stack
> if I donate it?
>
> Aslak
>
>>
>> On Jan 3, 2008 6:09 PM, aslak hellesoy <aslak.hellesoy at gmail.com>  
>> wrote:
>>> On Jan 3, 2008 12:15 AM, aslak hellesoy <aslak.hellesoy at gmail.com>  
>>> wrote:
>>>> Hi,
>>>>
>>>> I'm digging into the RubyWorks stack and learning about monit and
>>>> runit. From what I understand, monit is capable of starting and
>>>> stopping processes and services, so I'm wondering:
>>>>
>>>> Why is runit used to start services when monit can also do it?
>>>>
>>>
>>> Just to clarify: It seems to me that mongrels can be kept alive  
>>> with:
>>>  monit -> mongrel_cluster
>>>
>>> Rubyworks is doing (as far as I can tell):
>>>  monit -> runit -> mongrel_cluster
>>>
>>> Why is this? I also came over a year old blog post that describes
>>> monit and mongrel_cluster without any runit:
>>> http://software.pmade.com/blogs/ramblings/2006/12/27/mongrel-cluster-and-monit
>>>
>>> Aslak
>>>
>>>
>>>> Aslak
>>>>
>>
>>> _______________________________________________
>>> Rubyworks-users mailing list
>>> Rubyworks-users at rubyforge.org
>>> http://rubyforge.org/mailman/listinfo/rubyworks-users
>>>
>>
>>
>>
>> --
>> Jeff Xiong
>> Software Journeyman - http://gigix.thoughtworkers.org
>> Open Source Contributor - http://stomperl.googlecode.com/
>> Technical Evangelist - http://www.infoq.com/cn/
>> _______________________________________________
>> Rubyworks-users mailing list
>> Rubyworks-users at rubyforge.org
>> http://rubyforge.org/mailman/listinfo/rubyworks-users
>>
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users

From gigix1980 at gmail.com  Thu Jan  3 09:06:08 2008
From: gigix1980 at gmail.com (Jeff Xiong)
Date: Thu, 3 Jan 2008 22:06:08 +0800
Subject: [Rubyworks-users] why runit?
In-Reply-To: <8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
Message-ID: <20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>

Alexey, I remember that we talked about something like "Cap-like
installation" (forgive my poor memory if I'm wrong) a few months ago.
How's it going now? Aslak's idea sounds pretty cool to me.

On Jan 3, 2008 9:25 PM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> On Jan 3, 2008 1:33 PM, Jeff Xiong <gigix1980 at gmail.com> wrote:
> > Monit starts mongrel and haproxy processes in daemon mode. That means
> > you have to manage all those PID files, otherwise you may get some
> > zombie PID files and cannot restart mongrel/haproxy any more. That's
> > why we added runit: it starts those processes as its sub-processes and
> > doesn't need any PID file.
> >
> > So we just use monit to "monitor" mongrel/haproxy processes, and use
> > runit to "run it". It sounds reasonable, doesn't it?
> >
>
> (Reposting - you should set the REPLY-TO header for this list to
> rubyworks-users at rubyforge.org)
>
> Jeff, Alex,
>
> I thought it was the daemon/pid issue - just wasn't sure. Thanks for
> clarifying this.
>
> I've set up a clean ubuntu/rubyworks box at our intranet where we plan
> to host a dozen or so Rails apps with very low traffic.
> I'm thinking haproxy is not needed and one mongrel per app is probably enough.
>
> I'm going to use monit/runit (and vlad) to get these apps running.
>
> I'll probably make some vlad (Rake) plugins that will:
>
> * Generate and upload a new runit config for a mongrel
> * Generate and upload a new monit config for a mongrel
> * Generate and upload a new Apache config that will set up proxying
> via a virtual host
> * Plus a standard way to redeploy applications and restart them with monit
>
> Is this something that you'd be interested in including in the stack
> if I donate it?
>
> Aslak
>
> >
>
> > On Jan 3, 2008 6:09 PM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > > On Jan 3, 2008 12:15 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > > > Hi,
> > > >
> > > > I'm digging into the RubyWorks stack and learning about monit and
> > > > runit. From what I understand, monit is capable of starting and
> > > > stopping processes and services, so I'm wondering:
> > > >
> > > > Why is runit used to start services when monit can also do it?
> > > >
> > >
> > > Just to clarify: It seems to me that mongrels can be kept alive with:
> > >   monit -> mongrel_cluster
> > >
> > > Rubyworks is doing (as far as I can tell):
> > >   monit -> runit -> mongrel_cluster
> > >
> > > Why is this? I also came over a year old blog post that describes
> > > monit and mongrel_cluster without any runit:
> > > http://software.pmade.com/blogs/ramblings/2006/12/27/mongrel-cluster-and-monit
> > >
> > > Aslak
> > >
> > >
> > > > Aslak
> > > >
> >
> > > _______________________________________________
> > > Rubyworks-users mailing list
> > > Rubyworks-users at rubyforge.org
> > > http://rubyforge.org/mailman/listinfo/rubyworks-users
> > >
> >
> >
> >
> > --
> > Jeff Xiong
> > Software Journeyman - http://gigix.thoughtworkers.org
> > Open Source Contributor - http://stomperl.googlecode.com/
> > Technical Evangelist - http://www.infoq.com/cn/
> > _______________________________________________
> > Rubyworks-users mailing list
> > Rubyworks-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rubyworks-users
> >
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users
>



-- 
Jeff Xiong
Software Journeyman - http://gigix.thoughtworkers.org
Open Source Contributor - http://stomperl.googlecode.com/
Technical Evangelist - http://www.infoq.com/cn/

From aslak.hellesoy at gmail.com  Thu Jan  3 10:12:13 2008
From: aslak.hellesoy at gmail.com (aslak hellesoy)
Date: Thu, 3 Jan 2008 16:12:13 +0100
Subject: [Rubyworks-users] why runit?
In-Reply-To: <20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
	<20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
Message-ID: <8d961d900801030712i7b85dc27r5f6aa46029c17a4a@mail.gmail.com>

On Jan 3, 2008 3:06 PM, Jeff Xiong <gigix1980 at gmail.com> wrote:
> Alexey, I remember that we talked about something like "Cap-like
> installation" (forgive my poor memory if I'm wrong) a few months ago.
> How's it going now? Aslak's idea sounds pretty cool to me.
>

Thanks :-)

Here is what I'm thinking: I'd love for all of the rubyworks
configuration for an app to live under RAILS_ROOT/config/deploy,
checked into the SCM. There should be an easy way to generate them
too.

config/deploy/apache.conf # virtual host goes here
config/deploy/monit.conf
config/deploy/haproxy.conf
config/deploy/mongrel_[port-port+n].conf
config/deploy/service/mongrel_[port-port+n]/run # runit starter scripts

Logs would go in the standard Rails app log directory.
The config/deploy/monit.conf file would be included in  /etc/monit/monitrc
The runit starter scripts could be symlinked to /var/service unless
there is a simpler way to run them where they are
Finally each app could live under /var/www/rails/[appname]
I'd try to keep things decoupled from Rails as much as possible.

So I'm planning to create something that can generate these files
(probably based on rubigen), as well as a vlad plugin that will make
it easy to deploy and restart a RubyWorks app based on this
configuration. Someone else can do the cap equivalent :-)

WDYT?

This is more like the "client/admin" side of RubyWorks, and maybe
doesn't belong in RubyWorks proper. Should I set up a separate project
for it?

Aslak

>
> On Jan 3, 2008 9:25 PM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > On Jan 3, 2008 1:33 PM, Jeff Xiong <gigix1980 at gmail.com> wrote:
> > > Monit starts mongrel and haproxy processes in daemon mode. That means
> > > you have to manage all those PID files, otherwise you may get some
> > > zombie PID files and cannot restart mongrel/haproxy any more. That's
> > > why we added runit: it starts those processes as its sub-processes and
> > > doesn't need any PID file.
> > >
> > > So we just use monit to "monitor" mongrel/haproxy processes, and use
> > > runit to "run it". It sounds reasonable, doesn't it?
> > >
> >
> > (Reposting - you should set the REPLY-TO header for this list to
> > rubyworks-users at rubyforge.org)
> >
> > Jeff, Alex,
> >
> > I thought it was the daemon/pid issue - just wasn't sure. Thanks for
> > clarifying this.
> >
> > I've set up a clean ubuntu/rubyworks box at our intranet where we plan
> > to host a dozen or so Rails apps with very low traffic.
> > I'm thinking haproxy is not needed and one mongrel per app is probably enough.
> >
> > I'm going to use monit/runit (and vlad) to get these apps running.
> >
> > I'll probably make some vlad (Rake) plugins that will:
> >
> > * Generate and upload a new runit config for a mongrel
> > * Generate and upload a new monit config for a mongrel
> > * Generate and upload a new Apache config that will set up proxying
> > via a virtual host
> > * Plus a standard way to redeploy applications and restart them with monit
> >
> > Is this something that you'd be interested in including in the stack
> > if I donate it?
> >
> > Aslak
> >
> > >
> >
> > > On Jan 3, 2008 6:09 PM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > > > On Jan 3, 2008 12:15 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > > > > Hi,
> > > > >
> > > > > I'm digging into the RubyWorks stack and learning about monit and
> > > > > runit. From what I understand, monit is capable of starting and
> > > > > stopping processes and services, so I'm wondering:
> > > > >
> > > > > Why is runit used to start services when monit can also do it?
> > > > >
> > > >
> > > > Just to clarify: It seems to me that mongrels can be kept alive with:
> > > >   monit -> mongrel_cluster
> > > >
> > > > Rubyworks is doing (as far as I can tell):
> > > >   monit -> runit -> mongrel_cluster
> > > >
> > > > Why is this? I also came over a year old blog post that describes
> > > > monit and mongrel_cluster without any runit:
> > > > http://software.pmade.com/blogs/ramblings/2006/12/27/mongrel-cluster-and-monit
> > > >
> > > > Aslak
> > > >
> > > >
> > > > > Aslak
> > > > >
> > >
> > > > _______________________________________________
> > > > Rubyworks-users mailing list
> > > > Rubyworks-users at rubyforge.org
> > > > http://rubyforge.org/mailman/listinfo/rubyworks-users
> > > >
> > >
> > >
> > >
> > > --
> > > Jeff Xiong
> > > Software Journeyman - http://gigix.thoughtworkers.org
> > > Open Source Contributor - http://stomperl.googlecode.com/
> > > Technical Evangelist - http://www.infoq.com/cn/
> > > _______________________________________________
> > > Rubyworks-users mailing list
> > > Rubyworks-users at rubyforge.org
> > > http://rubyforge.org/mailman/listinfo/rubyworks-users
> > >
> > _______________________________________________
> > Rubyworks-users mailing list
> > Rubyworks-users at rubyforge.org
> > http://rubyforge.org/mailman/listinfo/rubyworks-users
> >
>
>
>
> --
> Jeff Xiong
> Software Journeyman - http://gigix.thoughtworkers.org
> Open Source Contributor - http://stomperl.googlecode.com/
> Technical Evangelist - http://www.infoq.com/cn/
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users
>

From alexey.verkhovsky at gmail.com  Thu Jan  3 11:01:18 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Thu, 3 Jan 2008 09:01:18 -0700
Subject: [Rubyworks-users] why runit?
In-Reply-To: <3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
	<20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
	<8d961d900801030712i7b85dc27r5f6aa46029c17a4a@mail.gmail.com>
	<3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
Message-ID: <3945c4270801030801l1152d7b3t5ba98a844585fbfc@mail.gmail.com>

On Jan 3, 2008 8:12 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> Here is what I'm thinking: I'd love for all of the rubyworks
> configuration for an app to live under RAILS_ROOT/config/deploy,
> checked into the SCM. There should be an easy way to generate them too.

There is definitely lack of flexibility in the cluster configuration,
which begs to be addressed.

I haven't given much thought to the details of it yet, but what you
propose (generating config files from some higher-level configuration
data, and placing them under ./config/deploy) is along the lines of
what I was thinking about.

Two questions:
1. Can we do that, but still keep the default case simple?
2. How to handle upgrades of the rubyworks package? Say, if there is a
change that requires re-generating config.

> This is more like the "client/admin" side of RubyWorks, and maybe
> doesn't belong in RubyWorks proper. Should I set up a separate project for it?

I hope not! :)

> each app could live under /var/www/rails/[appname]

>From a practical point of view, /var/www is static content in many
default Apache setups. From a theoretical one, /var is not where
application code is supposed to go as per FHS
[http://www.pathname.com/fhs/]. Hence the original choice of
/usr/rails. Nothing wrong with /usr/rails/[appname] (other than
backwards compatibility issues, maybe).

--

Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From alexey.verkhovsky at gmail.com  Thu Jan  3 11:36:31 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Thu, 3 Jan 2008 09:36:31 -0700
Subject: [Rubyworks-users] capistrano recipes ... rubyworks_deployer?
In-Reply-To: <ce1ffa990801021213y6153da90w6eb8dd42cbe2b7d0@mail.gmail.com>
References: <ce1ffa990801021213y6153da90w6eb8dd42cbe2b7d0@mail.gmail.com>
Message-ID: <3945c4270801030836o598f6b94r157a1956b3d3407e@mail.gmail.com>

On Jan 2, 2008 1:13 PM, John D. Hume <duelin.markers at gmail.com> wrote:
> Did we miss anything?

Not that I know of. :)
I've done Cap deployment a couple of times, but we haven't generalized it yet.

> What's the story on [ruyworks-deployer] plugin [in rubyworks trunk]?  Is it usable?

Probably not. However, rubyworks_deployer/recipes/rubyworks.rb is a
good starting point for your own Cap recipe.

-- 
Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From rubyworks at mlecza.newnetco.com  Thu Jan  3 13:14:14 2008
From: rubyworks at mlecza.newnetco.com (homerlex)
Date: Thu, 3 Jan 2008 13:14:14 -0500
Subject: [Rubyworks-users] Service Folder Creation
Message-ID: <200801031814.m03IELta028182@smtp.amsnet.net>

Hello,

 

 Just started playing around with the RubyWorks Production Stack and I have
a question about the folders in /var/service

 

It looks like mongrel_300x folders were created (perhaps during the
install).  What if I want to add more mongrels?  Do I need to manually
created more of these service folders (and edit the run script) or is there
something that is supposed to generate these folders for me?  I tried adding
a few more mongrels and unless I manually create these, I can't get them to
start up.

 

Thanks in advance.

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/rubyworks-users/attachments/20080103/c5ef773a/attachment.html 

From alexey.verkhovsky at gmail.com  Thu Jan  3 14:55:05 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Thu, 3 Jan 2008 12:55:05 -0700
Subject: [Rubyworks-users] Service Folder Creation
In-Reply-To: <200801031814.m03IELta028182@smtp.amsnet.net>
References: <200801031814.m03IELta028182@smtp.amsnet.net>
Message-ID: <3945c4270801031155t27c8a4b9x2139db4442ef46d@mail.gmail.com>

On Jan 3, 2008 11:14 AM, homerlex <rubyworks at mlecza.newnetco.com> wrote:
> It looks like mongrel_300x folders were created (perhaps during the
> install).  What if I want to add more mongrels?  Do I need to manually
> created more of these service folders (and edit the run script)

Yes. And edit Monit and HAProxy configurations.

> or is there something that is supposed to generate these folders for me?

As we discussed in another thread today, there should be, but there isn't.

-- 
Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From CWathing at thoughtworks.com  Thu Jan  3 17:00:41 2008
From: CWathing at thoughtworks.com (Chad Wathington)
Date: Thu, 3 Jan 2008 16:00:41 -0600
Subject: [Rubyworks-users] Chad Wathington is out of the office.
Message-ID: <OF2C27EE6A.BE423497-ON862573C5.0078E9B3-862573C5.0078E9B3@thoughtworks.com>


I will be out of the office starting  01/01/2008 and will not return until
01/11/2008.

I will check e-mail and voicemail throughout the day.


From mail at tsfsc.com  Fri Jan  4 05:34:44 2008
From: mail at tsfsc.com (J.F.Sebastian)
Date: Fri, 4 Jan 2008 02:34:44 -0800 (PST)
Subject: [Rubyworks-users] step-by-step instructions for rubyworks setup
	(and radiant on top)
Message-ID: <52173.91.11.235.154.1199442884.squirrel@webmail.tsfsc.com>

Hello everybody,

because I have to setup rubyworks installations again and again, I made a
step-by-step instruction list for setting up debian, mysql and rubyworks
(in a vmware) and how to install radiant on top of this.

I hope you like it:
Installation of debian, mysql and rubyworks (in a vmware)
http://www.tsfsc.com/blog/?p=6

Installation of radiant on top
http://www.tsfsc.com/blog/?p=6

Best regards,
J.F.

p.s. feel free to comment and suggest improvements.


From aslak.hellesoy at gmail.com  Fri Jan  4 11:54:29 2008
From: aslak.hellesoy at gmail.com (aslak hellesoy)
Date: Fri, 4 Jan 2008 17:54:29 +0100
Subject: [Rubyworks-users] why runit?
In-Reply-To: <3945c4270801030801l1152d7b3t5ba98a844585fbfc@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
	<20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
	<8d961d900801030712i7b85dc27r5f6aa46029c17a4a@mail.gmail.com>
	<3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
	<3945c4270801030801l1152d7b3t5ba98a844585fbfc@mail.gmail.com>
Message-ID: <8d961d900801040854q644165d7ua8b2eda35ce25c6a@mail.gmail.com>

On Jan 3, 2008 5:01 PM, Alexey Verkhovsky <alexey.verkhovsky at gmail.com> wrote:
>
> On Jan 3, 2008 8:12 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > Here is what I'm thinking: I'd love for all of the rubyworks
> > configuration for an app to live under RAILS_ROOT/config/deploy,
> > checked into the SCM. There should be an easy way to generate them too.
>
> There is definitely lack of flexibility in the cluster configuration,
> which begs to be addressed.
>
> I haven't given much thought to the details of it yet, but what you
> propose (generating config files from some higher-level configuration
> data, and placing them under ./config/deploy) is along the lines of
> what I was thinking about.
>
> Two questions:
> 1. Can we do that, but still keep the default case simple?
> 2. How to handle upgrades of the rubyworks package? Say, if there is a
> change that requires re-generating config.
>
> > This is more like the "client/admin" side of RubyWorks, and maybe
> > doesn't belong in RubyWorks proper. Should I set up a separate project for it?
>
> I hope not! :)
>
> > each app could live under /var/www/rails/[appname]
>
> >From a practical point of view, /var/www is static content in many
> default Apache setups. From a theoretical one, /var is not where
> application code is supposed to go as per FHS
> [http://www.pathname.com/fhs/]. Hence the original choice of
> /usr/rails. Nothing wrong with /usr/rails/[appname] (other than
> backwards compatibility issues, maybe).
>

Shouldn't it be /usr/local/webapps/[appname] then?
http://www.pathname.com/fhs/pub/fhs-2.3.html#THEUSRHIERARCHY

Aslak

(please fix the reply-to header to go to this list)

> --
>
> Alexey Verkhovsky
> CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
> RubyWorks [http://rubyworks.thoughtworks.com]
>
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users
>

From ian.w.white at gmail.com  Fri Jan  4 12:13:36 2008
From: ian.w.white at gmail.com (Ian White)
Date: Fri, 4 Jan 2008 17:13:36 +0000
Subject: [Rubyworks-users] why runit?
In-Reply-To: <8d961d900801040854q644165d7ua8b2eda35ce25c6a@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
	<20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
	<8d961d900801030712i7b85dc27r5f6aa46029c17a4a@mail.gmail.com>
	<3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
	<3945c4270801030801l1152d7b3t5ba98a844585fbfc@mail.gmail.com>
	<8d961d900801040854q644165d7ua8b2eda35ce25c6a@mail.gmail.com>
Message-ID: <f66db9cb0801040913p6d3be720t630aa6d3fdcf3f09@mail.gmail.com>

As far as I am aware /usr shouldn't have app specific data.  Distros
can (have license to) blow away stuff in /usr without asking
questions.

(from the page you just mentioned) "Any information that is
host-specific or varies with time is stored elsewhere."

/srv is what I use for this stuff, so what about /srv/webapps/[APPNAME]
It is also recommended in the document Aslak mentioned

"/srv contains site-specific data which is served by this system"

on one of my machines:

/srv/svn
/srv/mephisto
/srv/trac
/srv/lighttpd

When I want to do a backup of all server related stuff it's /srv I grab.

Just my ?0.02

Cheers,
Ian


-- 
Argument from Design--We build web applications
Western House 239 Western Road Sheffield S10 1LE UK
Mobile: +44 (0)797 4678409 | Office: +44 (0)114 2667712
<http://www.ardes.com/> | <http://blog.ardes.com/ian>

From aslak.hellesoy at gmail.com  Fri Jan  4 13:17:52 2008
From: aslak.hellesoy at gmail.com (aslak hellesoy)
Date: Fri, 4 Jan 2008 19:17:52 +0100
Subject: [Rubyworks-users] why runit?
In-Reply-To: <3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
	<20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
	<8d961d900801030712i7b85dc27r5f6aa46029c17a4a@mail.gmail.com>
	<3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
Message-ID: <8d961d900801041017n2927e7e9w9acc2c33a80358bc@mail.gmail.com>

On Jan 3, 2008 5:00 PM, Alexey Verkhovsky <alexey.verkhovsky at gmail.com> wrote:
> On Jan 3, 2008 8:12 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > Here is what I'm thinking: I'd love for all of the rubyworks
> > configuration for an app to live under RAILS_ROOT/config/deploy,
> > checked into the SCM. There should be an easy way to generate them too.
>
> There is definitely lack of flexibility in the cluster configuration,
> which begs to be addressed.
>
> I haven't given much thought to the details of it yet, but what you
> propose (generating config files from some higher-level configuration
> data, and placing them under ./config/deploy) is along the lines of
> what I was thinking about.
>
> Two questions:
> 1. Can we do that, but still keep the default case simple?

What is the default case?

> 2. How to handle upgrades of the rubyworks package? Say, if there is a
> change that requires re-generating config.
>

I'm not sure yet. Please take a look at what I have done so far:

https://boss.bekk.no/repos/projects/rubyworks_app/

It's currently undocumented - I'm only playing with the generator
(based on rubigen). Quick intro

svn co https://boss.bekk.no/repos/projects/rubyworks_app/rubyworks_app
cd rubyworks_app
rake install_gem # you may have to install some gems for this to work
cd [your_rails_app]
rubyworks_app --haproxy-port 3010 --ports [3011, 3012, 3013]
Then look at the generated files under config/deploy (I haven't done
haproxy.conf yet - nor apache.conf)

The next step is to generate the vlad/cap file that will let you say:
rake deploy
or
cap deploy

and have everything updated on the server, create symlinks and bounce
monit (and maybe apache)

I may donate this stuff to RubyWorks when it matures a little.
Feedback is welcome.

Aslak

> > This is more like the "client/admin" side of RubyWorks, and maybe
> > doesn't belong in RubyWorks proper. Should I set up a separate project for it?
>
> I hope not! :)
>
> > each app could live under /var/www/rails/[appname]
>
> From a practical point of view, /var/www is static content in many
> default Apache setups. From a theoretical one, /var is not where
> application code is supposed to go as per FHS
> [http://www.pathname.com/fhs/]. Hence the original choice of
> /usr/rails. Nothing wrong with /usr/rails/[appname] (other than
> backwards compatibility issues, maybe).
>
> --
>
> Alexey Verkhovsky
> CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
> RubyWorks [http://rubyworks.thoughtworks.com]
>

From alexey.verkhovsky at gmail.com  Fri Jan  4 13:26:52 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Fri, 4 Jan 2008 11:26:52 -0700
Subject: [Rubyworks-users] why runit?
In-Reply-To: <8d961d900801041017n2927e7e9w9acc2c33a80358bc@mail.gmail.com>
References: <8d961d900801021515g791fd861pc532075abbf5ca64@mail.gmail.com>
	<8d961d900801030209h6f080ae6pade6acf8ca44a2fb@mail.gmail.com>
	<20fd9dc50801030433y3e324493rf84d033c1f0242e4@mail.gmail.com>
	<8d961d900801030525k1b751a75rec87148cee79da09@mail.gmail.com>
	<20fd9dc50801030606o5003277r83aaeda33aa0e526@mail.gmail.com>
	<8d961d900801030712i7b85dc27r5f6aa46029c17a4a@mail.gmail.com>
	<3945c4270801030800x5833698eje6f199ecdaf778f0@mail.gmail.com>
	<8d961d900801041017n2927e7e9w9acc2c33a80358bc@mail.gmail.com>
Message-ID: <3945c4270801041026t62c84bd5w47f7913271ce63d@mail.gmail.com>

On Jan 4, 2008 11:17 AM, aslak hellesoy <aslak.hellesoy at gmail.com> wrote:
> > 1. Can we do that, but still keep the default case simple?
> What is the default case?
Single, low volume app on a dedicated [virtual] server. Basically,
where you are OK with the defaults and don't care to reconfigure
anything.

> I'm not sure yet. Please take a look at what I have done so far:
Will do.

-- 
Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From bigbet at 126.com  Fri Jan  4 19:29:27 2008
From: bigbet at 126.com (bigbet)
Date: Sat, 5 Jan 2008 08:29:27 +0800
Subject: [Rubyworks-users] fffffffff
Message-ID: <200801050829267967170@126.com>






bigbet
2008-01-05
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/rubyworks-users/attachments/20080105/8f571130/attachment.html 

From rubyworks at mlecza.newnetco.com  Wed Jan  9 09:41:20 2008
From: rubyworks at mlecza.newnetco.com (homerlex)
Date: Wed, 9 Jan 2008 09:41:20 -0500
Subject: [Rubyworks-users] Why is ruby mysql driver being used?
Message-ID: <200801091441.m09EfQg5007020@smtp.amsnet.net>

I installed the Production Stack on Centos 5.  I was having some perf
problems with my queries on this server - also getting sporadic "Lost
Connection" errors.

 

I put a debug in the mysql_adapter to see the version of the driver that was
being used and it shows 4.0-ruby-0.2.6-plus-changes.  Obviously it using the
ruby mysql driver instead of the mysql gem.  

 

Anyone have an idea why this is?  When I do a gem list -local I do see the
mysql 2.7 gem installed.

 

To verify that it wasn't just something with my custom rails app, I also
dupliated using the default rails app that Production Stack installs (by
creating a controller that references a Model class.

 

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/rubyworks-users/attachments/20080109/83d3d45e/attachment.html 

From alexey.verkhovsky at gmail.com  Wed Jan  9 09:53:22 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Wed, 9 Jan 2008 07:53:22 -0700
Subject: [Rubyworks-users] Why is ruby mysql driver being used?
In-Reply-To: <200801091441.m09EfQg5007020@smtp.amsnet.net>
References: <200801091441.m09EfQg5007020@smtp.amsnet.net>
Message-ID: <3945c4270801090653n2521cd9cq435f113fa6f7d6eb@mail.gmail.com>

On Jan 9, 2008 7:41 AM, homerlex <rubyworks at mlecza.newnetco.com> wrote:
> I put a debug in the mysql_adapter to see the version of the driver that was
> being used and it shows 4.0-ruby-0.2.6-plus-changes.  Obviously it using the
> ruby mysql driver instead of the mysql gem.

What happens if you open IRB and put the following two lines in it:

require 'rubygems'
require 'mysql'

-- 
Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From rubyworks at mlecza.newnetco.com  Wed Jan  9 10:24:43 2008
From: rubyworks at mlecza.newnetco.com (homerlex)
Date: Wed, 9 Jan 2008 10:24:43 -0500
Subject: [Rubyworks-users] Using a custom GEM PATH in Production Stack
Message-ID: <200801091524.m09FOe98008248@smtp.amsnet.net>

 

What is the recommended way to have my Rails app use  a custom Gem Path?

 

My server may eventually have several different applications and they will
need to use different gems.

 

Regards

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://rubyforge.org/pipermail/rubyworks-users/attachments/20080109/fab89c7b/attachment-0001.html 

From duelin.markers at gmail.com  Wed Jan  9 22:55:48 2008
From: duelin.markers at gmail.com (John D. Hume)
Date: Wed, 9 Jan 2008 22:55:48 -0500
Subject: [Rubyworks-users] Using a custom GEM PATH in Production Stack
In-Reply-To: <200801091524.m09FOe98008248@smtp.amsnet.net>
References: <200801091524.m09FOe98008248@smtp.amsnet.net>
Message-ID: <ce1ffa990801091955l54769232wbee86dc5b33933e9@mail.gmail.com>

Personally I'd recommend (wherever possible) unpacking gems into the
vendor directory of the app that uses them, even if there's only going
to be one application on your server.  Anything you can do to reduce
your app's dependency on a special environment and move closer to
just-checkout-and-it-works is a good thing.

-hume

On Jan 9, 2008 10:24 AM, homerlex <rubyworks at mlecza.newnetco.com> wrote:
>
>
>
>
>
>
> What is the recommended way to have my Rails app use  a custom Gem Path?
>
>
>
> My server may eventually have several different applications and they will
> need to use different gems.
>
>
>
> Regards
>
>
> _______________________________________________
> Rubyworks-users mailing list
> Rubyworks-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/rubyworks-users
>
>

From alexey.verkhovsky at gmail.com  Fri Jan 11 11:23:00 2008
From: alexey.verkhovsky at gmail.com (Alexey Verkhovsky)
Date: Fri, 11 Jan 2008 09:23:00 -0700
Subject: [Rubyworks-users] Using a custom GEM PATH in Production Stack
In-Reply-To: <ce1ffa990801091955l54769232wbee86dc5b33933e9@mail.gmail.com>
References: <200801091524.m09FOe98008248@smtp.amsnet.net>
	<ce1ffa990801091955l54769232wbee86dc5b33933e9@mail.gmail.com>
Message-ID: <3945c4270801110823i3ae9734bw1bd1d9d64e434172@mail.gmail.com>

On Jan 9, 2008 8:55 PM, John D. Hume <duelin.markers at gmail.com> wrote:
> Personally I'd recommend (wherever possible) unpacking gems into the
> vendor directory of the app

+1 on "vendor everything" recommendation.

-- 
Alexey Verkhovsky
CruiseControl.rb [http://cruisecontrolrb.thoughtworks.com]
RubyWorks [http://rubyworks.thoughtworks.com]

From zhcchz at gmail.com  Wed Jan 16 05:02:21 2008
From: zhcchz at gmail.com (Chris Stevenson)
Date: Wed, 16 Jan 2008 18:02:21 +0800
Subject: [Rubyworks-users] Installing on Ubuntu 7.10
Message-ID: <64766f910801160202o150c1264h6ac69c6133ef2c88@mail.gmail.com>

I managed to install on Ubuntu 7.10 today, and it worked fine except
for one small issue. When the rails application is installed the log
files have the wrong permissions. If I start the application from the
command prompt I get an error saying that the log file cannot be
written to, and that it should be chmod 0666.

If I do this then everything works:

chmod 0666 log
chmod 0666 log/*

chris