Mongrel users - Mar 2007 - FreeBSD 6.1 mongrel_cluster and procfs message

I''m running the latest mongrel (1.0.1) and the latest mongrel_cluster  
(1.0.1.1) on FreeBSD 6.1.

When I use capistrano to try to stop or restart my cluster I get this  
(partial) output:

ps: Process environment requires procfs(5)

And my Mongrel processes haven''t stopped.

Any ideas?

Thanks,

Jamie

Jamie Orchard-Hays wrote:
> I''m running the latest mongrel (1.0.1) and the latest
mongrel_cluster
> (1.0.1.1) on FreeBSD 6.1.
> 
> When I use capistrano to try to stop or restart my cluster I get this  
> (partial) output:
> 
> ps: Process environment requires procfs(5)
> 
> And my Mongrel processes haven''t stopped.
> 
> Any ideas?
The message means that you need the procfs virtual filesystem mounted. 
Normally you do not need this under FreeBSD.

I use Mongrel 1.0.1 + mongrel_cluster (0.2.1) on FreeBSD 6.2 without any 
glitches, so maybe mongrel_cluster 1.0.1.1 introduces some 
Linux-specific ps syntax that requires procfs.

I will try to upgrade my mongrel_cluster version and see if I get the 
same message.
> 
> Thanks,
> 
> Jamie
Jonathan

-- 
Jonathan Weiss
http://blog.innerewut.de

Thanks. I''d found a bit about that after I sent the message. Things  
worked fine with 0.2.1--except that mongrel_cluster''s restart did not  
work properly if nothing was running, which cost me an hour a day  
earlier before I realized what was wrong. That''s was my motivation to  
upgrade.

If you have any simple instructions about how to get this working on  
FreeBSD, I''ll be very grateful!

Jamie

On Mar 4, 2007, at 4:17 AM, Jonathan Weiss wrote:
> The message means that you need the procfs virtual filesystem mounted.
> Normally you do not need this under FreeBSD.
>
> I use Mongrel 1.0.1 + mongrel_cluster (0.2.1) on FreeBSD 6.2  
> without any
> glitches, so maybe mongrel_cluster 1.0.1.1 introduces some
> Linux-specific ps syntax that requires procfs.
>
> I will try to upgrade my mongrel_cluster version and see if I get the
> same message.

Kirk, I followed the instructions on the following link, but no love.  
I seem to have mounted fine, but I''m getting the same error reported  
back from mongrel_cluster_ctl.

http://os.newsforge.com/os/06/03/22/1531252.shtml?tid=8&tid=2

Jamie

On Mar 4, 2007, at 10:44 AM, Jamie Orchard-Hays wrote:
> Thanks. I''d found a bit about that after I sent the message.
Things
> worked fine with 0.2.1--except that mongrel_cluster''s restart did
not
> work properly if nothing was running, which cost me an hour a day
> earlier before I realized what was wrong. That''s was my motivation
to
> upgrade.
>
> If you have any simple instructions about how to get this working on
> FreeBSD, I''ll be very grateful!
>
> Jamie
>
> On Mar 4, 2007, at 4:17 AM, Jonathan Weiss wrote:
>
>> The message means that you need the procfs virtual filesystem  
>> mounted.
>> Normally you do not need this under FreeBSD.
>>
>> I use Mongrel 1.0.1 + mongrel_cluster (0.2.1) on FreeBSD 6.2
>> without any
>> glitches, so maybe mongrel_cluster 1.0.1.1 introduces some
>> Linux-specific ps syntax that requires procfs.
>>
>> I will try to upgrade my mongrel_cluster version and see if I get the
>> same message.
>
> _______________________________________________
> Mongrel-users mailing list
> Mongrel-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/mongrel-users

On 3/4/07, Jamie Orchard-Hays <jamie at dangosaur.us>
wrote:
> Kirk, I followed the instructions on the following link, but no love.
> I seem to have mounted fine, but I''m getting the same error
reported
> back from mongrel_cluster_ctl.
>
> http://os.newsforge.com/os/06/03/22/1531252.shtml?tid=8&tid=2
>
> Jamie
>
It is a ps command problem.  I ran into the same issue when testing
locally on OS X.  Options are to wait for another *pre-release* of
mongrel_cluster or making the following mod to line 164 of init.rb in
the check_process method:

old:   ps_output = `ps -o args= -p #{pid}`
new: ps_output = `ps -o command= -p #{pid}`

Pat

Patrick, I assume you mean the gem file on the remote server. This  
didn''t change the error I get, or cause restart to work.

Thanks,
Jamie

On Mar 4, 2007, at 11:36 AM, Patrick Berry wrote:
> It is a ps command problem.  I ran into the same issue when testing
> locally on OS X.  Options are to wait for another *pre-release* of
> mongrel_cluster or making the following mod to line 164 of init.rb in
> the check_process method:
>
> old:   ps_output = `ps -o args= -p #{pid}`
> new: ps_output = `ps -o command= -p #{pid}`
>
> Pat

This ''ps'' syntax is really annoying. If my score-keeping is
correct:

Linux and FreeBSD seem to support both ''args'' and
''command''.
Darwin/MacOSX doesn''t support ''args''.
Solaris doesn''t support ''command''.

I guess I need to resort to platform checks. I was hoping to avoid this 
by using the ''standard format specifiers''. The great thing
about
standards is that there are so many to choose from.

Bradley

Patrick Berry wrote:
> On 3/4/07, Jamie Orchard-Hays <jamie at dangosaur.us> wrote:
>> Kirk, I followed the instructions on the following link, but no love.
>> I seem to have mounted fine, but I''m getting the same error
reported
>> back from mongrel_cluster_ctl.
>>
>> http://os.newsforge.com/os/06/03/22/1531252.shtml?tid=8&tid=2
>>
>> Jamie
>>
> 
> It is a ps command problem.  I ran into the same issue when testing
> locally on OS X.  Options are to wait for another *pre-release* of
> mongrel_cluster or making the following mod to line 164 of init.rb in
> the check_process method:
> 
> old:   ps_output = `ps -o args= -p #{pid}`
> new: ps_output = `ps -o command= -p #{pid}`
> 
> Pat
> _______________________________________________
> Mongrel-users mailing list
> Mongrel-users at rubyforge.org
> http://rubyforge.org/mailman/listinfo/mongrel-users