Plugins » Katello

Cloned from https://bugzilla.redhat.com/show_bug.cgi?id=2168865

Description of problem:

One has the ability to create a Registry Service account and generate a token related to the same to further use them as shared credentials but the generated token could be quite long.

While for a container\docker type repo, The "Upstream Password" field accepts that Token, The "Upstream Authentication Token" field rejects the same with error

An error occurred saving the Repository: PG::StringDataRightTruncation: ERROR: value too long for type character varying(1024)

Version-Release number of selected component (if applicable):

Any version that accepts the upstream auth token, likely 3.18+

How reproducible:

Easily and Always

Steps to Reproduce:
1. Create a Katello server

2. Import a manifest with some valid subs

3. Go to https://access.redhat.com/terms-based-registry/#/accounts and create a new service account and token.

4. Discovery "rhel8/toolbox" docker type repo using the exact same username and token from UI Content --> Products page and create the repo

5. Sync the repo

6. Edit the "Upstream Authorization" field and try to put the token inside ""Upstream Authentication Token" field -> Save

7. Try to remove the token from there and put it in "Upstream Password" field and save

8. Re-sync the repo.

Actual results:

Step 5: Sync Success ( as the token gets saved as password )

Step 6: Error:

An error occurred saving the Repository: PG::StringDataRightTruncation: ERROR: value too long for type character varying(1024)

Step 7 and 8: Successfully saved and repo can be synced

The reason seems to be this i.e. upstream_password column has a record type set to Text but upstream_authentication_token is of varchar(1024)

1. echo "\d+ katello_root_repositories;" | su - postgres -c "psql foreman" | egrep "Column|upstream" | awk -F'|' '{print $1,$2,$4}'
   Column Type Nullable
   docker_upstream_name character varying(255)
   upstream_username character varying(255)
   upstream_password text
   upstream_authentication_token character varying(1024)

Expected results:

A) As tokens are expected to be longer than normal passwords, set "upstream_authentication_token" type to "text" or increase the length for varchar

B) If the token can be used in both upstream_password as well as upstream_authentication_token fields, Then just have one field present.

Additional info:

NA