https://www.puppet.com/community logo
Join Slack
Powered by
# puppet-enterprise
  • d
    yeah, followed that exact guide, but not sure how to check 😉
  • b
    I never set that, always xms/xmx directly
  • b
    how does your jvm params look? in /etc/sysconfig/pe-puppetserver.conf if you're on rhel-like systems
  • b
    take the max memory value, divide by jruby instances, should be 2048mb
  • d
    hmmm, that file doesn’t seem to exist on the compiler 🤔
  • b
    on which OS are you
  • d
    OracleLinux, so rhel-based
  • d
    OracleLinux, so rhel-based
  • b
    it's /etc/sysconfig/pe-puppetserver
  • b
    without the .conf
  • b
    check for the JAVA_ARGS
  • d
    Copy code
    # Modify this if you'd like to change the memory allocation, enable JMX, etc
JAVA_ARGS="-Xmx6144m -Xms6144m -Xss2m -Djava.io.tmpdir=/opt/puppetlabs/server/apps/puppetserver/tmp -XX:ReservedCodeCacheSize=384m -XX:+UseParallelGC -Xlog:gc*:file=/var/log/puppetlabs/puppetserver/puppetserver_gc.log::filecount=16,filesize=65536 -Djdk.tls.ephemeralDHKeySize=2048"

# Modify this as you would JAVA_ARGS but for non-service related subcommands
JAVA_ARGS_CLI="${JAVA_ARGS_CLI:-}"
  • b
    that looks wrong
  • d
    in what sense?
  • b
    ah I mixed up the -xss with -xms and thought -xms is listed twice
  • s
    This message was deleted.
    m
    b
    • 3
    • 3
  • s
    This message was deleted.
    v
    r
    t
    • 4
    • 5
  • b
    i usually have something like this:
    Copy code
    puppet_enterprise::profile::master::java_args: {"Xms":"16g","Xmx":"16g"}
puppet_enterprise::master::puppetserver::jruby_max_active_instances: 8
puppet_enterprise::master::puppetserver::reserved_code_cache: 2048m
  • d
    both on primary as compilers?
  • b
    my rule oh thumb: is it a big env? 2gig memory per jruby instance. one jruby instance per cpu core. on small envs 1gig mem per jruby. always 2gig reserved code cache. valid for FOSS/PE primary/PE compiler. if puppetdb has a huge workload, do not reserve all cpu cores for puppetserver. run puppetdb/puppetserver with java 11. use latest postgresql you can get, ensure jruby9k is used.
    👍 1
  • b
    my rule oh thumb: is it a big env? 2gig memory per jruby instance. one jruby instance per cpu core. on small envs 1gig mem per jruby. always 2gig reserved code cache. valid for FOSS/PE primary/PE compiler. if puppetdb has a huge workload, do not reserve all cpu cores for puppetserver. run puppetdb/puppetserver with java 11. use latest postgresql you can get, ensure jruby9k is used.
  • m
    2mb is the default
  • v
    that puts proper settings into pe.conf
  • s
    Hello Community, I have a question: I am planning to move a PE installation to another Server. I used the node_manager module to dump the current configuration. I replaced the relationship between roles and users with the name to get rid of the IDs. What options do I have to deal with the rbac_role permissions, especially with the instance IDs? Do I replace all IDs with * to be able to apply the permissions to the new server?
  • v
    well, you can use pltraining-rbac in similar matter, but you can backup/restore all of them with standard PE backup tools
  • v
  • s
    @vchepkov I will have a look at this
    👍 1
  • s
    This message was deleted.
    n
    n
    • 3
    • 5
  • b
    Hi people, reading https://puppet.com/docs/pe/2021.5/code_mgr_customizing.html#config_post_env_hooks it mentions the following example for a code manager hook:
    Copy code
    puppet_enterprise::master::code_manager::post_environment_hooks: 
  - url: '<https://console.yourorg.com:4433/classifier-api/v1/update-classes>'    
    use-client-ssl: true
    is there any reason why this isn't enabled by default?
  • s
    This message was deleted.
    s
    • 2
    • 1
1...567...73Latest