In January, QRIScloud operational staff will be performing required software upgrades to the Ceph servers that provide the QRIScloud / NeCTAR and QRIScloud / RDS volume storage pools. The schedule is as follows:
- 2017-01-04 - 09:00 to 17:00 - Upgrading the NeCTAR Ceph cluster from "Hammer" to "Jewel".
- 2017-01-09 - 09:00 to 17:00 - Upgrading the RDS Ceph cluster from "Firefly" to "Hammer".
- 2017-01-16 - 09:00 to 17:00 - Upgrading the RDS Ceph cluster from "Hammer" to "Jewel".
This work is not anticipated to lead to significant service interruptions or outages. However, as with all work of this nature, it is possible for something to go wrong, leading (in a worst-case scenario) to loss of data.
We therefore advise QRIScloud volume storage users as follows:
- Make sure that you have the appropriate backup procedures in place for the data in your volumes.
- Check that your backups are working properly. In the (unlikely) event of data loss caused by a failed Ceph upgrade, your backups could be the only way to get your data back.
- If your volume storage dependent systems experience operational problems during the advertised maintenance windows, please let us know promptly.
UPDATE - 2017-01-04 - 16:55 - The maintenance work scheduled for today was completed without any incidents.
UPDATE - 2017-01-09 - 16:30 - The maintenance work scheduled for today could not be done due to a couple of more urgent incidents. We will attempt to do it tomorrow.
UPDATE - 2017-01-11 - 11:55 - The maintenance work rescheduled for yesterday was completed. The final work originally scheduled for the 16th has been brought forward and is being done today.
UPDATE - 2017-01-12 - 09:25 - The final part of maintenance work was completed without incident.