Self-hosting
Troubleshooting
When an installation does not start, the licence is not renewed or uploads fail.
A service does not start
The logs show the reason:
docker compose logs cms --tail 100
kubectl -n ddm logs deploy/cms --tail 100
DATABASE_URLor password wrong: the password in.envmust be the one Postgres was set up with on the first start.- Migration failed: the CMS then does not start, and the database stays as it was. Send us the logs.
- Images cannot be pulled: is the server signed in to
registry.gitlab.com? A deploy token we have revoked no longer pulls anything.
The licence is not renewed
Settings › Licence shows the last error.
| Error | Cause |
|---|---|
fetch failed, ENOTFOUND, timeout | The server cannot reach cms.documentdistributionmanager.com. Open outgoing HTTPS to it, or set HTTPS_PROXY. |
OTHER_INSTALLATION | The licence belongs to another installation. Send us the Installation ID. |
UNKNOWN_LICENCE, INVALID_LICENCE | The licence is not (or no longer) on record with us. Write to us. |
notYetValid | The server's clock is wrong. Set up NTP. |
Uploads fail
- Large files: does your reverse proxy let requests through up to
MAX_UPLOAD_BYTES? nginx needsclient_max_body_size 1g;, ingress-nginx theproxy-body-sizeannotation, whichddm.yamlalready sets. - No previews: is the render worker running? It needs access to the same volume or the same bucket as the CMS.
Sign-in to the reader app does not work
If the sign-in works in the CMS but not in the reader app, AUTH_COOKIE_DOMAIN is wrong: it must be the domain both host names share.
Help from us
In DDM, choose Suggest a feature or Report a bug. The report reaches us directly, with the technical details we need to track down the problem.