ManageIQ on RHEL8 | Puppet exec `find_spec_for_exe': can't find gem manageiq-appliance_console error

[needle2k]

hi,

i´ve created a puppet class to install ManageIQ Oparin on RHEL8 via Satellite so far everthing works fine. If i config it via appliance_console on console afterwards.

i wanted to extend my class to call appliance_console_cli via puppet exec

$command.each |$i, $cmd| {
      exec {"db_create_${i}":
        command => "appliance_console_cli ${cmd}",
        path          => $::path,
        onlyif        => "test ! -f /root/db_created_${i}.sem",
        require     => Package['manageiq-gemset'],
      }
      # create semaphor file
      file { "/root/db_created_${i}.sem":
        ensure    => 'present',
        subscribe => Exec["db_create_${i}"],
      }
    }

from the output it seems that i might miss some Parameters eg. environment?

Notice: /Stage[main]/manageiq_server::Manageiq/Exec[db_create_0]/returns: /usr/share/rubygems/rubygems.rb:278:in `find_spec_for_exe': can't find gem manageiq-appliance_console (>= 0.a) with executable appliance_console_cli (Gem::GemNotFoundException)
Notice: /Stage[main]/manageiq_server::Manageiq/Exec[db_create_0]/returns: 	from /usr/share/rubygems/rubygems.rb:297:in `activate_bin_path'
Notice: /Stage[main]/manageiq_server::Manageiq/Exec[db_create_0]/returns: 	from appliance_console_cli:25:in `<main>'
Error: 'appliance_console_cli --hostname localhost --username username --password password --dbname vmdb_production --region 0 --port 5432 --dbdisk=/dev/vdb' returned 1 instead of one of [0]
Error: /Stage[main]/manageiq_server::Manageiq/Exec[db_create_0]/returns: change from 'notrun' to ['0'] failed: 'appliance_console_cli --hostname localhost --username username --password password --dbname vmdb_production --region 0 --port 5432 --dbdisk=/dev/vdb' returned 1 instead of one of [0] (corrective)

any help to lift me on track would be appreciated, i´m not that familiar with ruby

[kbrock]

That sounds like it is not setting the path variables correctly for ruby.

I wonder if sourcing /etc/default/evm would get the environment setup in a way that fixes things.

You can probably see that the manageiq-*@.service files load EnvironmentFile=/etc/default/manageiq*.properties. Don't think these would affect you but wanted to throw that out there.

[needle2k]

Hey kbrock,

sourcing directly /etc/default/evm didn´t help on the "appliance_console_cli" part, i´ve added an additional 'environment' paramter to my puppet exec command with values taken from your hint about the manageiq*.properties files

$command.each |$i, $cmd| {
      exec {"db_create_${i}":
        command     => "appliance_console_cli ${cmd}",
        environment => ['BUNDLE_GEMFILE=/var/www/miq/vmdb/Gemfile','GEM_HOME=/opt/manageiq/manageiq-gemset','GEM_PATH=/opt/manageiq/manageiq-gemset:/usr/share/gems:/usr/local/share/gems','TERM=xterm'],
        path        => $::path,
        onlyif      => "test ! -f /root/db_created_${i}.sem",
        require     => Package['manageiq-gemset'],
      }
      # create semaphor file
      file { "/root/db_created_${i}.sem":
        ensure    => 'present',
        subscribe => Exec["db_create_${i}"],
      }
    }

afterwards it nagged about the db related ENV Parameters so i ended in "source /etc/default/evm" and done my appliance_console_cli exec part, now it works!

thx ;)