"foreman-maintain backup" fails on CentOS8 / Foreman 2.1.0rc* - Multiple Issues

Problem: foreman-maintain backup does not function on CentOS8

Expected outcome: A clean run of the foreman-maintain backup command

Foreman and Proxy versions: 2.1.0rc*

Foreman and Proxy plugin versions: 2.1.0rc*

Distribution and version: CentOS 8.2

Other relevant data:

While I realize the documentation states that you need Katello (Content Management) installed to backup the foreman instance, that seems odd to me, as there is data to be backed up in a generic foreman instance, so I pushed onwards.

I tried (as root):

foreman-maintain backup offline --skip-pulp-content --features all --include-db-dumps /tmp/backups
...snipped...
--------------------------------------------------------------------------------
Backup config files:                                                  [FAIL]
undefined method `[]' for nil:NilClass
--------------------------------------------------------------------------------

That failed, this has been brought up in other support posts without any one answering! Example: Backup-Restore Strategy in Foreman - In my testing, it also fails on CentOS 7. :frowning:

So lets whitelist it as suggested and move on:

foreman-maintain backup offline --whitelist="backup-config-files" --skip-pulp-content --features all --include-db-dumps /tmp/backups
...snipped...
--------------------------------------------------------------------------------
Backup config files:                                                  [SKIPPED]
--------------------------------------------------------------------------------
Backup Foreman DB offline: 
- Collecting data from /var/opt/rh/rh-postgresql12/lib/pgsql/data/    [FAIL]    
No such file or directory @ dir_chdir - /var/opt/rh/rh-postgresql12/lib/pgsql/data/
--------------------------------------------------------------------------------

That doesn’t look good, and in fact it is a hard-coded database location for foreman >= 2.0, which is valid for CentOS 7, but not valid for CentOS 8. The following shows some details and the problematic code:

# dnf list installed | grep maintain
rubygem-foreman_maintain.noarch   1:0.6.5-1.el8   @foreman

# ps -ef | grep postmaster
postgres  ...snipped... /usr/bin/postmaster -D /var/lib/pgsql/data

# pwd
/usr/share/gems/gems/foreman_maintain-0.6.5

# grep -r -C 5 /var/opt/rh
...snipped...
definitions/features/foreman_database.rb-  def config_files
definitions/features/foreman_database.rb-    if check_min_version('foreman', '2.0')
definitions/features/foreman_database.rb:      ['/var/opt/rh/rh-postgresql12/lib/pgsql/data/postgresql.conf']
definitions/features/foreman_database.rb-    else
definitions/features/foreman_database.rb-      ['/var/lib/pgsql/data/postgresql.conf']
definitions/features/foreman_database.rb-    end
definitions/features/foreman_database.rb-  end
definitions/features/foreman_database.rb-
--
lib/foreman_maintain/concerns/base_database.rb-module ForemanMaintain
lib/foreman_maintain/concerns/base_database.rb-  module Concerns
lib/foreman_maintain/concerns/base_database.rb-    module BaseDatabase
lib/foreman_maintain/concerns/base_database.rb-      def data_dir
lib/foreman_maintain/concerns/base_database.rb-        if check_min_version('foreman', '2.0')
lib/foreman_maintain/concerns/base_database.rb:          '/var/opt/rh/rh-postgresql12/lib/pgsql/data/'
lib/foreman_maintain/concerns/base_database.rb-        else
lib/foreman_maintain/concerns/base_database.rb-          '/var/lib/pgsql/data/'
lib/foreman_maintain/concerns/base_database.rb-        end
lib/foreman_maintain/concerns/base_database.rb-      end
lib/foreman_maintain/concerns/base_database.rb-

As a new foreman user, the inability to backup a basic foreman instance (let alone, restore it), does not seem great given that foreman-maintain is an installed command by default. And while I ignored a prerequisite, surely backing up a foreman+smart-proxy instance (c.f. foreman+smart-proxy+katello+pulp instance), should be a subset of the later. This was all done on the same CentOS 8 VM documented in my previous post: Problem with Foreman 2.1rc2 with Remote Execution on CentOS 8 - #33 by illumino

Is a basic backup/restore mechanism for generic foreman instances in the works? Surely it must be simplier or a subset of the +Katello+Pulp configuration.

PS: The database backup did complete on CentOS 7 and restart foreman. But as stated above the config files backup failed!

for the first issue, I guess it’s: Bug #30166: Backup fails with undefined method `[]' for nil:NilClass - Foreman Maintain - Foreman
for the wrong path, I’ve opened Bug #30323: assumes SCL paths for databases on EL8 - Foreman Maintain - Foreman

Hi,

I also had this issue on CentOS 8 but with foreman-2.4.0-0.4.rc2 and rubygem-foreman_maintain-0.7.4-1

As a workaround, you can create the missing directory and create a symlink

install -d /var/opt/rh/rh-postgresql12/lib && ln -s /var/lib/pgsql /var/opt/rh/rh-postgresql12/lib/pgsql

Hope it helps.

Mike