Failled to add new Datanode

Hello !

  • Are now running the same graylog server and data node versions?
    Yes, absolutely… Not!
    Due to all my tests and snapshot restorations, the nodes and server were not on the same version.
    I’ve now upgraded everything to version 6.2.2-1.

Have you deleted content of mongodb before starting all services?
Yes, here’s what I did:

  • Shut down the data node, Graylog server, and MongoDB

  • On the data node: deleted data and config files

  • On the server: deleted the contents of /var/lib/mongodb

  • Rebooted all VMs

  • Launched the preflight web interface

  • Created the CA

  • Configured the renewal policy

  • Provisioned the certificates

  • Prayed

  • no changes

Are you using any custom JVM or those bundled with graylog server and data node? Any JAVA_HOME or OPENSEARCH_JAVA_HOME env properties?
No custom JVM.
I just installed Debian 12 and followed the Graylog installation instructions from the official documentation.

A few moments later… i’ve got a clue.

  1. When i run only one datanode, it seems to work fine :

  1. I add one node, click “restart the configuration”, still looks good :

  1. I add the third one…

  1. I delete the first two…

  1. I turn the first two back on…

  2. Click “Restart Configuration”

  3. And… still the same issue as at the beginning of the week!


Maintenance operations like recalculate or rotate are painless.

If i clear everythings and restart the preflight…

The only thing that always works is with one node without any problem !