At about 2:20pm our cluster died. It was back up and running by 2:30pm.
One of our cluster computers died, taking email and home directories down with it. We discovered an issue which was causing the other half of the cluster to pause and not respond to email and home directory requests. That problem should be fixed and when 1/2 the cluster dies in the future, fewer services will stop working as a result. Idealy, no one should notice if the redundancy of the cluster is as it should be.
Leave a Reply