Puppet users - Apr 2013 - Flush provider - Differentiating between new resource and modification?

Morning all

I''m working on converting some of my NetApp providers to prefetch/flush
style to try and optimize performance. 

I''ve hit an issue on my Netapp_user provider, around handling resource 
creation versus resource modification? 
What''s the easiest way to differentiate? 

Current code is here: 
https://github.com/fatmcgav/fatmcgav-netapp/commit/66092978f4182c5474a60011db99ee2e3e12e689

Any tips appreciated. 

Regards
Gavin 

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to puppet-users+unsubscribe@googlegroups.com.
To post to this group, send email to puppet-users@googlegroups.com.
Visit this group at http://groups.google.com/group/puppet-users?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

On Fri, 5 Apr 2013 00:57:32 -0700 (PDT)
Gavin Williams <fatmcgav@gmail.com> wrote:
> Morning all
> 
> I''m working on converting some of my NetApp providers to
> prefetch/flush style to try and optimize performance. 
> 
> I''ve hit an issue on my Netapp_user provider, around handling
> resource creation versus resource modification? 
> What''s the easiest way to differentiate? 
> 
> Current code is here: 
>
https://github.com/fatmcgav/fatmcgav-netapp/commit/66092978f4182c5474a60011db99ee2e3e12e689
> 
> Any tips appreciated. 
> 
> Regards
> Gavin 
> 
There is no way to check *why* the flush method was called, you just now
that at least one property has been updated. You do not see if `ensure`
updated or let''s say `passmaxage`. Does this actually cause problems?

One thing I''ve spotted is that your create method does update the
@property_hash[:ensure] value but no other value. This seems to be
wrong because if the resource was absent before, @property_hash is
initally an empty hash. Because when `ensure` changes no other
properties are synced you don''t have the desired values of all the
other
properties available in the `flush` method. So your `create` method
should propably look like
    
    def create
      resource.class.validproperties.each do |property|
        if value = resource.should(property)
          @property_hash[property] = value
        end
      end
    end

-Stefan

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to puppet-users+unsubscribe@googlegroups.com.
To post to this group, send email to puppet-users@googlegroups.com.
Visit this group at http://groups.google.com/group/puppet-users?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.

On Saturday, April 6, 2013 5:22:03 PM UTC-5, Stefan Schulte
wrote:
>
> On Fri, 5 Apr 2013 00:57:32 -0700 (PDT) 
> Gavin Williams <fatm...@gmail.com <javascript:>> wrote: 
>
> > Morning all 
> > 
> > I''m working on converting some of my NetApp providers to 
> > prefetch/flush style to try and optimize performance. 
> > 
> > I''ve hit an issue on my Netapp_user provider, around handling
> > resource creation versus resource modification? 
> > What''s the easiest way to differentiate? 
> > 
> > Current code is here: 
> > 
>
https://github.com/fatmcgav/fatmcgav-netapp/commit/66092978f4182c5474a60011db99ee2e3e12e689
> > 
> > Any tips appreciated. 
> > 
> > Regards 
> > Gavin 
> > 
>
> There is no way to check *why* the flush method was called, you just now 
> that at least one property has been updated. You do not see if `ensure` 
> updated or let''s say `passmaxage`. Does this actually cause
problems?
>

Your prefetch() method determines which target resources already exist as 
part of its normal job.  It could flag those resources via some internal 
variable of the resource, or perhaps even in the @property_hash.  You could 
even formalize that by creating a read-only property for it, if that makes 
you more comfortable.  At flush() time you can assume that resources that 
are flagged already exist, and those that are not, don''t.  There is a
small
window for collision if physical resources of the target type can be 
created / destroyed by a different process while Puppet is running, but 
that issue exists in any case.

If you were not doing prefetching then the provider could check the 
existence of the physical resource as the first step of flushing.  Indeed, 
it would need to do so, one way or another, to correctly apply the 
resource''s ''ensure'' parameter.  Your provider can
work this way despite
prefetching if you don''t want to mark prefetched resources. 


John

-- 
You received this message because you are subscribed to the Google Groups
"Puppet Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to puppet-users+unsubscribe@googlegroups.com.
To post to this group, send email to puppet-users@googlegroups.com.
Visit this group at http://groups.google.com/group/puppet-users?hl=en.
For more options, visit https://groups.google.com/groups/opt_out.