Created attachment 676276 [details] engine logs Description of problem: --------------------------------------- When installation of a host fails to happen and it is in the 'Install Failed' state, if it is moved to 'Maintenance' mode and then activated, it goes to non-responsive and then immediately disappears from the Console with the following message in the Events log - "Server <IP> was removed from Cluster <cluster-name> from gluster CLI. Now removed in engine DB as well." Version-Release number of selected component (if applicable): Red Hat Storage Console Version: 2.1-0.qa4d.el6rhs How reproducible: Frequently Steps to Reproduce: 1. Add a host to a cluster and let it be 'UP' 2. Add another host such that installation for it fails and it goes to 'Install Failed' state. 3. Move the host (for which install failed) to 'Maintenance' mode and then try to Activate it. Actual results: The host disappears from the Console with the above described message seen in the Events log. Expected results: The host should not be removed, it should remain in the 'Non-responsive' state. Additional info:
This is not the right way to recover from a failed installation. If installation fails, you should see an alert message that looks like: "Host installation failed. Fix installation issues and try to <a>Re-Install</a>" You should just click on the "Re-Install" link to retry the installation. As long as the host is in "Install Failed" status, or any status other than Up/Down/Maintenance, it will not be removed automatically by the sync job. In this case, it is getting removed because you move it to maintenance, and glusterfs reports that it is not part of the cluster.