Closed akariv closed 5 years ago
the relevant error is in the secrets
initContainer logs -
Initializing datastore DB demo-instance-61-datastore on ckan-cloud.cokcn6iyrwpx.us-west-2.rds.amazonaws.com
CREATE ROLE
ALTER DATABASE
WARNING: no privileges were granted for "spatial_ref_sys"
WARNING: no privileges were granted for "geography_columns"
WARNING: no privileges were granted for "geometry_columns"
WARNING: no privileges were granted for "raster_columns"
WARNING: no privileges were granted for "raster_overviews"
ALTER DEFAULT PRIVILEGES
CREATE VIEW
GRANT
CREATE FUNCTION
ERROR: permission denied for relation spatial_ref_sys
CONTEXT: SQL statement "CREATE TRIGGER zfulltext BEFORE INSERT OR UPDATE ON spatial_ref_sys FOR EACH ROW EXECUTE PROCEDURE populate_full_text_trigger();"
PL/pgSQL function inline_code_block line 3 at EXECUTE
Datastore DB initialization failed
failed to initialize, waiting 1 hour before exiting
I'll have a look tomorrow
there are some problems with the way we are waiting for CKAN instance to be ready:
to solve this problems I'm implementing a cca-operator instance-health
command which will report accurate status and health for the instance, will use it in create-instance script
Any idea why initContainer failed in the first place?
On Mon, Dec 10, 2018 at 10:19 AM Ori Hoch notifications@github.com wrote:
there are some problems with the way we are waiting for CKAN instance to be ready:
- the initContainer restarts on failure (after 1 hour) - after restart it detects the secrets already created and then completes successfully, hiding the error and allowing CKAN to start
- we are checking CKAN ready status in create instance script to continue but it could be reported as ready and fail a few seconds later on init scripts
to solve this problems I'm implementing a cca-operator instance-health command which will report accurate status and health for the instance, will use it in create-instance script
— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/ViderumGlobal/ckan-cloud-helm/issues/5#issuecomment-445728573, or mute the thread https://github.com/notifications/unsubscribe-auth/AAQMdSEs45kYTqte9_3DwIN_0I4pnMi1ks5u3hkLgaJpZM4ZKP0X .
i was lucky enough to get it before the 1 hour ended.. so the log I pasted above is the first error
create-instance log:
Inspecting the ckan pod: