Puppet users - Mar 2013 - File resources: using source much slower than content

I''m moving from version 2.7.19 to 3.1.0 and am running into an issue
with
file resources.

All of them using a "source => ''...''" attribute
to pull their contents from
the master are taking about 5 seconds longer than they used to.  These are 
generally very small files which used to transfer very quickly.  This 
happens with both webrick and passenger.  If I turn them into templates 
which don''t actually do any substitutions and use a "content =>
template(...)", then they go back to taking a fraction of a second each.

Given that the evaluation times for each are around 5.05 seconds when they 
used to be around 0.05 seconds, it looks like they''re waiting 5 seconds
for
something to time out, but I can''t figure out what.

Has anyone seen anything like this?

-- 
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 Monday, March 11, 2013 4:40:09 PM UTC-5, Tim Moorhouse
wrote:
>
> I''m moving from version 2.7.19 to 3.1.0 and am running into an
issue with
> file resources.
>
> All of them using a "source => ''...''"
attribute to pull their contents
> from the master are taking about 5 seconds longer than they used to.  These
> are generally very small files which used to transfer very quickly.  This 
> happens with both webrick and passenger.  If I turn them into templates 
> which don''t actually do any substitutions and use a "content
=>
> template(...)", then they go back to taking a fraction of a second
each.
>
> Given that the evaluation times for each are around 5.05 seconds when they 
> used to be around 0.05 seconds, it looks like they''re waiting 5
seconds for
> something to time out, but I can''t figure out what.
>
> Has anyone seen anything like this?
>

It is natural that ''source'' would be slower than
''content'', even much
slower.  With ''content'', the target file contents are embedded
in the
catalog, whereas with ''source'', the agent makes a separate
request to the
server for each file.  Network requests are relatively slow in most 
environments.  I would recommend using ''content'' for small
files wherever
possible.  To that end, it may be easier to use the file() function instead 
the template() function when the content you want to serve is completely 
static.

With that said, I don''t know why the timing would be so strikingly 
different in 3.1 than it was in 2.7, if everything else is the same.  If no 
one has an answer to that, and you want to troubleshoot it, then I would 
suggest starting by identifying whether the delay is at the master or at 
the agent.  You should be able to tell by checking the file service 
request-to-response turnaround times in the master''s logs.


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.

Hi Tim,

On Monday, March 11, 2013, Tim Moorhouse wrote:
> I''m moving from version 2.7.19 to 3.1.0 and am running into an
issue with
> file resources.
>
> All of them using a "source => ''...''"
attribute to pull their contents
> from the master are taking about 5 seconds longer than they used to.  These
> are generally very small files which used to transfer very quickly.  This
> happens with both webrick and passenger.  If I turn them into templates
> which don''t actually do any substitutions and use a "content
=>
> template(...)", then they go back to taking a fraction of a second
each.
>
> Given that the evaluation times for each are around 5.05 seconds when they
> used to be around 0.05 seconds, it looks like they''re waiting 5
seconds for
> something to time out, but I can''t figure out what.
>
Sounds like reverse DNS lookups are failing:
https://projects.puppetlabs.com/issues/18573<https://projects.puppetlabs.com/issues/18573#change-87186>.
Can you update with your webrick or passenger info and ruby version?

> Has anyone seen anything like this?
>
> --
> 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 <javascript:_e({},
> ''cvml'',
''puppet-users%2Bunsubscribe@googlegroups.com'');>.
> To post to this group, send email to
puppet-users@googlegroups.com<javascript:_e({}, ''cvml'',
''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.
>
>
>

-- 
Josh Cooper
Developer, Puppet Labs

-- 
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.