Why do we see messages such as "Unable to self-heal permissions/ownership of '/' (possible split-brain)" in the logs after one node crashed ?
Issue
- After a gluster node have to be forcefully restarted due to a hang or a crash, the log files are filled up with the following messages :
Unable to self-heal permissions/ownership of '/' (possible split-brain). Please fix the file on all backend volumes
- Why is this so ?
Environment
-
Glusterfs 3.2.2 (Legacy Gluster)
-
Red Hat Storage Software Appliance 3.2
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.