X-Git-Url: https://git.arvados.org/arvados.git/blobdiff_plain/0f644e242ef37c911ad3dc25aca8135c339de349..0fa894e23fb824ea9b5222d808f3c205651d3170:/build/rails-package-scripts/README.md

diff --git a/build/rails-package-scripts/README.md b/build/rails-package-scripts/README.md
index 0d720bde14..6ac2539f8e 100644
--- a/build/rails-package-scripts/README.md
+++ b/build/rails-package-scripts/README.md
@@ -12,7 +12,6 @@ Since our build process is a tower of shell scripts, concatenating files seemed
 
 postinst.sh lets the early parts define a few hooks to control behavior:
 
-* After it installs the core configuration files (database.yml, application.yml, and production.rb) to /etc/arvados/server, it calls setup_extra_conffiles.  By default this is a noop function (in step2.sh).  API server defines this to set up the old omniauth.rb conffile.
+* After it installs the core configuration files (database.yml, application.yml, and production.rb) to /etc/arvados/server, it calls setup_extra_conffiles.  By default this is a noop function (in step2.sh).
 * Before it restarts nginx, it calls setup_before_nginx_restart.  By default this is a noop function (in step2.sh).  API server defines this to set up the internal git repository, if necessary.
-* $RAILSPKG_DATABASE_LOAD_TASK defines the Rake task to load the database.  API server uses db:structure:load.  SSO server uses db:schema:load.  Workbench doesn't set this, which causes the postinst to skip all database work.
-* If $RAILSPKG_SUPPORTS_CONFIG_CHECK != 1, it won't run the config:check rake task.  SSO clears this flag (it doesn't have that task code).
+* $RAILSPKG_DATABASE_LOAD_TASK defines the Rake task to load the database.  API server uses db:structure:load.  Workbench doesn't set this, which causes the postinst to skip all database work.