Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Graylog startup behaviour #1136
Currently when a server node starts and finds issues with the Elasticsearch cluster it refuses to complete the startup process - with the introduction of the journaling subsystem in 1.0 I find this behaviour a bit problematic since it could mean loosing messages.
I would like to suggest that server nodes always start regardless of the ES cluster status (and thus can buffer messages in it's journal) - the web interface could then inform the user that ES has a problem that needs investigating to avoid confusion.