> On 13 Dec 2017, at 20:48, n.bele...@gaijin.ru wrote: > > You are managing lots of files on your agent. > How many are managed? > grep file /opt/puppetlabs/puppet/cache/state/resources.txt | wc -l > > > grep -c file /opt/puppetlabs/puppet/cache/state/resources.txt > 223
That is not many. > > Do you manage directories using recursive => true? > > No, we do not use it at all. Good. > > Are most files templates (using content => epp(…)) or static files (using > source => ‘puppet:///…’)? > > > 30% templates\70% static. In this case I would analyze the network latency between the master and the agent. Every file resource with source attribute gets fetched from master with a new https + client-cert connection. If network latency is the cause, we usually try to remove source and use content with template on all files. Yes, this increases compile times on the master and the catalog gets larger. But for the benefit that the agent has all file content and never must initialize a call to the master. > > Is it possible to determine what file resources require a lot of time? Hard to say. Maybe enabling agent debug mode prior first puppet run. But that will produce a big debug file to parse. Either run puppet agent —test —debug or add debug=true to puppet.conf agent section. Please remeber to remove the debug after you have your analysis. Are there big files you sync to the agents? Files with several 100 MB? In this case it might be the generation of md5 sums. > > On the machine where we run agent 32 core and 24GB of memory. Enough for puppet agent ;-) -- 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/8D494A51-9A88-4396-9D2C-7CC301DFD6DA%40gmail.com. For more options, visit https://groups.google.com/d/optout.
