Hi,Â
Until yesterday, an agent on a node was applying the catalog very fast (about 5 seconds).
Yesterday, i've made  a definition that manages the fstab, using the module "herculesteam-augeasproviders_mounttab". The code appears to be very simple :
define linuxbasico::fstab(
 $mntpointsource,
 $mntpointdir,
 $mntpointtype,
 $mntpointoptions
) {
 file { "${mntpointdir}":
  ensure => directory,
 }
 ->
 mounttab { "${mntpointdir}":
  ensure  => present,
  device  => "${mntpointsource}",
  fstype  => "${mntpointtype}",
  options  => $mntpointoptions,
  dump   => "0",
  pass   => "0",
  provider => augeas,
 }
}
So, i've some tests to manage a mount point through this definition and everything worked fine. But, after a few more tests, the agent was taking too long to apply the catalog (more than 100 seconds). So, i removed the reference to my new definition and also removed the mounttab module, but the slowness persisted.
What i've noticed through debug messages is the following group of messages lots of times:
Debug: Failed to load library 'msgpack' for feature 'msgpack'
Debug: Puppet::Network::Format[msgpack]: feature msgpack is missing
Debug: file_metadata supports formats: pson yaml binary
Another detail, after the message:
Notice: Applied catalog in 104.19 seconds
the agent is still hanging for about 30 seconds.
I've added the following config on the puppetserver, but nothing changed
always_cache_features = true
Any tips?
Thanks in advance.
Ps: Debug log file is attached
Lauro