Starting on the 3rd of July 2018 and continuing until the 10th of July the NCI NeCTAR node will be performing a series of rolling reboots across all our hypervisors in order to install updated versions of the BIOS and firmware. This is required to provide proper protection against a number of security vulnerabilities.
For the updates to take effect the hypervisors need to be rebooted. During the week we will be rebooting blocks of ten to fifteen hypervisors each day, with the owners of affected instances being notified 24 hours in advance. We cannot specify exactly when instances may be affected, except to note that all reboots will happen during business hours.
In order to minimise the impact on users we will attempt to live-migrate active instances off the hypervisors that are being rebooted - these instances will not be impacted by the maintenance work. In some cases live migration may not be possible - in these cases the affected instance will be suspended before the node is rebooted, and resumed afterwards. Suspended instances will be unavailable for approximately ten minutes, but will not be rebooted - existing processes will remain, though network connections and interactive sessions will be dropped during the suspension.
Notifications will be sent out to all directly affected users ahead of the actual maintenance work.