This evening at 18:00, there will be a short outage of the NFS service that provides access to collections and other file system held in the main Polaris GPFS servers. The aim of the outage is to address an NFS performance issue by migrating NFS metadata to a dedicated device. The outage is expected to take less than 5 minutes.
We anticipate that people, applications and services that attempt to access GPFS via NFS will experience a pause of up to 5 minutes during the outage. Service should then resume automatically.
While we do not expect this to occur, it is possible that client systems might experience "NFS stale file handle" problems which might lead to user or application-level I/O errors. If this does occur, the solution is to manually unmounting and remounting the affected collection. If this doesn't solve the problem, please contact QRIScloud Support.
Systems / services that are likely to be impacted by this outage include:
- NeCTAR VMs that mount RDS collections via 10.255.122.70
- Access to RDS collections on GPFS via "ssh1", "ssh2", "data", "nextcloud" or "gridftp"
- Access to RDS collections on GPFS from the QRIScloud and RCC HPC systems; i.e. "awoonga", "flashlite" and "tinaroo".
UPDATE: The change was implemented by 18:10.