/scratch restored following outage this morning
Update, 11/1/2021 3pm: The /scratch
storage sub-system has been remounted across the cluster and the public gateways. There are a few custom gateways that will need to be rebooted and that will be coordinated with the respective groups.
One of the components for the /scratch
storage sub-system entered a bad state over the weekend. Our sysadmins were able to force it into a good state only to have it reoccur. The result is that some /scratch
users have experienced intermittent issues when accessing files.
We plan on rebooting the component in order to clear the issue. To make sure this doesn’t impact other parts of the system, we will be taking /scratch
offline for 1 hour at 11am this morning.