With a very small edge-case asterisk (the puppet Device support, which I've
never used), yes. The agent is a piece of software that runs (as a daemon,
via cron, manually, or via some other scheduling mechanism like
mcollective) on a host ("node" in Puppet nomenclature) and applies catalogs
to that one node.There's a good introduction to this at https://docs.puppetlabs.com/learning/agent_master_basic.html - I'd also encourage you to look through the rest of the Learning Puppet section on docs.puppetlabs.com. On Sat, Sep 27, 2014 at 12:23 PM, Dennis Gearon <[email protected]> wrote: > Thanks Jason. If I ever get to 1000 nodes, then I guess I'll have > something to worry about, then. LOL! One final question. > > An 'agent' is really a managed server instance, not a Puppet worker > managing something else, (as the word 'Agent' would suggest). So really, > Agents only apply catalogs to themselves, right? > > On Sat, Sep 27, 2014 at 9:07 AM, Jason Antman <[email protected]> > wrote: > >> Agents never control other agents. Aside from supporting technology >> (PuppetDB, an ENC if you have one, a database to back PuppetDB), yeah, a >> master and N agents is the gist of it. >> >> There are some docs online (see specifically the "Tuning and Scaling" >> section of the Puppet Documentation Index, >> https://docs.puppetlabs.com/puppet/#tuning-and-scaling ) about scaling, >> and a search of the mailing list archives (as well as Google - there have >> been quite a few blog posts on this) should help turn up other options and >> experiences. >> >> Off the top of my head, I believe there are generally two paradigms used: >> either clustering/HA for your masters, generally load-balanced from what I >> gather, or (what I do) splitting up agents between different masters (by >> location, or network, or dev/test/prod). My current infrastructure is made >> up of ~450 nodes which are served by three masters - dev, test/QA and prod. >> We run every 30 minutes, via mcollective with a set concurrency. The >> masters are VMs, running on the same physical hosts as PuppetDB and its >> PostgreSQL instance, so I'm quite confident that I could scale each one to >> a few thousand nodes before I have to worry about overloading one host. In >> addition, since we split dev, test/QA and prod masters, we can deploy >> module changes (and puppet/puppetdb/etc. upgrades) to the dev environment, >> validate there, promote to test/QA, validate there, and then promote to >> prod. >> >> -Jason >> >> On Sat, Sep 27, 2014 at 11:03 AM, Dennis Gearon <[email protected]> >> wrote: >> >>> Been burnnig up the keyboard and spewing packets to search for this >>> answer, but haven't seen it. >>> >>> From what I've read, there is only: >>> A/ A Puppet Master >>> B/ Infinite number of 'Agent' nodes. >>> >>> Is this right? >>> >>> Is there any other kinds of nodes? >>> >>> Do Agent nodes ever control other nodes? >>> >>> What happens when the puppet master gets overloaded, how do you cluster >>> or use a 'super' master to divide the load? >>> >>> Thanks for clearning this up for me :-) >>> >>> -- >>> 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 [email protected]. >>> To view this discussion on the web visit >>> https://groups.google.com/d/msgid/puppet-users/4c6dde8c-d82e-4159-85a0-6fe4cc7aeb04%40googlegroups.com >>> <https://groups.google.com/d/msgid/puppet-users/4c6dde8c-d82e-4159-85a0-6fe4cc7aeb04%40googlegroups.com?utm_medium=email&utm_source=footer> >>> . >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- >> 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/AdEjg1IBGUc/unsubscribe. >> To unsubscribe from this group and all its topics, send an email to >> [email protected]. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/puppet-users/CAFt4V4%3DkTtL1O8JNeKLidmLpym61q8DYdPmVy4rJTa%2BOJLV_0g%40mail.gmail.com >> <https://groups.google.com/d/msgid/puppet-users/CAFt4V4%3DkTtL1O8JNeKLidmLpym61q8DYdPmVy4rJTa%2BOJLV_0g%40mail.gmail.com?utm_medium=email&utm_source=footer> >> . >> >> For more options, visit https://groups.google.com/d/optout. >> > > -- > 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 [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/puppet-users/CAAJay%2BZWWCGneffuK0voo5Ummx8sJxY0GD1q7ioeuE8Wh9ee3Q%40mail.gmail.com > <https://groups.google.com/d/msgid/puppet-users/CAAJay%2BZWWCGneffuK0voo5Ummx8sJxY0GD1q7ioeuE8Wh9ee3Q%40mail.gmail.com?utm_medium=email&utm_source=footer> > . > > For more options, visit https://groups.google.com/d/optout. > -- 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 [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/puppet-users/CAFt4V4nwUGg%3DA4XQk7AC3psUFuQUQtNug6pLC68Hy6r8s2K7cA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
