On Wed, May 7, 2014 at 6:24 AM, Antoine Cotten <tonio.cot...@gmail.com>wrote:
> Thanks Josh, > > This seems to work better indeed, although I can still observe a decrease > in performances. The config retrieval time is higher than with config-based > environments (9 sec -> 26 sec on average), and CPU-wise I can also notice > an increase but this is hard to quantify without prolonged monitoring, I'll > let it run for 24 extra hours before giving any conclusion. But, well, > nothing compared to v3.5.1 fortunately, so I believe this RC1 is promising! > > Feel free to ask if you need any data or extra feedback from me. > Thanks for testing this, Antoine! There is a time to live setting for the new environment cache which defaults to 5 seconds. It's 'environment_timeout' and you can set it in your puppet.conf, or per directory environment in it's environment.conf. I'm curious what, if any, change you see setting it to say '30s', and what change if you set it 'unlimited'. Would you have time to test that? thanks, Josh > -- > Toni > > > > > On Tuesday, May 6, 2014 10:55:24 PM UTC+2, Josh Partlow wrote: >> >> Hi Antoine, >> >> There is a Puppet 3.6.0-rc1 available in the pre-release repos which has >> environment cacheing that we think will take care of the slowdown you are >> seeing with directory environments. Would you be able to test that? (See >> Eric's 3.6.0-rc1 announcement for details on obtaining the packages) >> >> thanks, >> Josh >> >> On Tuesday, April 22, 2014 11:26:47 AM UTC-7, Eric Sorenson wrote: >>> >>> >>> >>> On Tuesday, April 22, 2014 5:25:55 AM UTC-7, Antoine Cotten wrote: >>>> >>>> Hi Eric, >>>> thank you for pointing me to this feature, and sorry for the answer >>>> time. >>>> >>>> I managed to outline the problematic parts in the Puppet run: the final >>>> POST method takes forever when Directory Environments is enabled (1) and >>>> the catalog compilation is overall much slower (1.3./4|5|6/) >>>> >>>> I triggered a run on the same node, with and without the feature >>>> enabled. Here are the corresponding profiling logs: >>>> >>>> >>> Thanks, this is great information. >>> >>> The reason the POST looks slow is because it's an aggregate of all the >>> steps that came before it, as you see the profile step numbers pop out a >>> level, that line aggregates the times of everything inside it. So the POST >>> is the aggregate of all the steps to compilation. >>> >>> I am looking at this more closely and I have opened https://tickets. >>> puppetlabs.com/browse/PUP-2337 to track it. >>> >>> --eric0 >>> >> -- > You received this message because you are subscribed to a topic in the > Google Groups "Puppet Users" group. > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/puppet-users/wzy8NPWauu4/unsubscribe. > To unsubscribe from this group and all its topics, send an email to > puppet-users+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/puppet-users/0008dd5d-fab6-4d01-b206-6a1994236b42%40googlegroups.com<https://groups.google.com/d/msgid/puppet-users/0008dd5d-fab6-4d01-b206-6a1994236b42%40googlegroups.com?utm_medium=email&utm_source=footer> > . > > For more options, visit https://groups.google.com/d/optout. > -- Josh Partlow jpart...@puppetlabs.com Developer, Puppet Labs Join us at PuppetConf 2014, September 22-24 in San Francisco - http://puppetconf.com Register by May 30th to take advantage of the Early Adopter discount —save $349! -- 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+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/CADxAQ5pxLHOSC9n6m3UTVGDX9HGY2KTXYnG1p1-qN9qcULLsnA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
