improve config options docs

config/galaxy.yml.sample

@@ -1511,11 +1511,14 @@ galaxy:
   # is 5MB, but as low as 1MB seems to be a reasonable size.
   #max_metadata_value_size: 5242880
 
-  # If (for example) you run on a cluster and your datasets (by default,
-  # database/files/) are mounted read-only, this option will override
-  # tool output paths to write outputs to the working directory instead,
-  # and the job manager will move the outputs to their proper place in
-  # the dataset directory on the Galaxy server after the job completes.
+  # This option will override tool output paths to write outputs to the
+  # job working directory (instead of to the file_path) and the job
+  # manager will move the outputs to their proper place in the dataset
+  # directory on the Galaxy server after the job completes. This is
+  # necessary (for example) if jobs run on a cluster and datasets can
+  # not be created by the user running the jobs (e.g. if the filesystem
+  # is mounted read-only or the jobs are run by a different user than
+  # the galaxy user).
   #outputs_to_working_directory: false
 
   # If your network filesystem's caching prevents the Galaxy server from
@@ -1547,27 +1550,27 @@ galaxy:
   #cleanup_job: always
 
   # When running DRMAA jobs as the Galaxy user
-  # (https://galaxyproject.org/admin/config/performance/cluster
-  # /#submitting-jobs-as-the-real-user) this script is used to run the
+  # (https://docs.galaxyproject.org/en/latest/admin/cluster.html
+  # #submitting-jobs-as-the-real-user) this script is used to run the
   # job script Galaxy generates for a tool execution.
   #drmaa_external_runjob_script: sudo -E scripts/drmaa_external_runner.py --assign_all_groups
 
   # When running DRMAA jobs as the Galaxy user
-  # (https://galaxyproject.org/admin/config/performance/cluster
-  # /#submitting-jobs-as-the-real-user) this script is used to kill such
+  # (https://docs.galaxyproject.org/en/latest/admin/cluster.html
+  # #submitting-jobs-as-the-real-user) this script is used to kill such
   # jobs by Galaxy (e.g. if the user cancels the job).
   #drmaa_external_killjob_script: sudo -E scripts/drmaa_external_killer.py
 
   # When running DRMAA jobs as the Galaxy user
-  # (https://galaxyproject.org/admin/config/performance/cluster
-  # /#submitting-jobs-as-the-real-user) this script is used transfer
+  # (https://docs.galaxyproject.org/en/latest/admin/cluster.html
+  # #submitting-jobs-as-the-real-user) this script is used transfer
   # permissions back and forth between the Galaxy user and the user that
   # is running the job.
   #external_chown_script: sudo -E scripts/external_chown_script.py
 
   # When running DRMAA jobs as the Galaxy user
-  # (https://galaxyproject.org/admin/config/performance/cluster
-  # /#submitting-jobs-as-the-real-user) Galaxy can extract the user name
+  # (https://docs.galaxyproject.org/en/latest/admin/cluster.html
+  # #submitting-jobs-as-the-real-user) Galaxy can extract the user name
   # from the email address (actually the local-part before the @) or the
   # username which are both stored in the Galaxy data base. The latter
   # option is particularly useful for installations that get the