Maintenance windows HPC Cloud

Current

Up & Running.

Future maintenance windows

Note that not every maintenance window will actually include that we will do maintenance. Please see the policy below for more information.

Status Date (start / end) comment
Scheduled 08-10-2019 / 09-10-2019  
Scheduled 12-11-2019 / 13-11-2019  
Scheduled 10-12-2019 / 11-12-2019  
Scheduled 14-01-2020 / 15-01-2020  
Scheduled 11-02-2020 / 12-02-2020  
Scheduled 10-03-2020 / 11-03-2020  
Scheduled 14-04-2020 / 15-04-2020  
Scheduled 12-05-2020 / 13-05-2020  
Scheduled 09-06-2020 / 10-06-2020  
Scheduled 14-07-2020 / 15-07-2020  
Scheduled 11-08-2020 / 12-08-2020  
Scheduled 08-09-2020 / 09-09-2020  
Scheduled 13-10-2020 / 14-10-2020  
Scheduled 10-11-2020 / 11-11-2020  
Scheduled 08-12-2020 / 09-12-2020  

Policy

Maintenance and changes can only be done during on of to the maintenance windows. This means that (part of) the service can be unavailable, for example the web UI.

All HPC Cloud users are notified at least one month in advance if:

If no changes are planned or only changes with some short interruptions expected (< 5 minutes), no notification will be send.

Exceptions

Exceptions to this policy are made in the following cases:

In these cases, the changes will be applied as soon as needed, even outside maintenance windows.

Past

(most recent first)

Status Date (start / end)  
Cancelled 10-09-2019 / 11-09-2019 Upgrade ONe to 5.8 and replace network components. Impact: service unavailable; all VMs need to go down.
Finished 31-07-2019 09:00 - 16:20 Unscheduled maintenance. Impact: UI/API unreachable; major network outage affecting many VMs; Ceph backend unreachable.
Finished 31-07-2019 07:00 - 09:00 Update network infrastructure components. Impact: short interruptions (+/- 1 min) for traffic can be expected; ‘lightpath’ link unavailable.
Finished 09-07-2019 13:00 - 15:24 Maintenance scheduled to improve service backend resilience. Impact: UI/API will not be available. VMs are not affected.
Finished 27-06-2019 20:03 - 20:10 Security maintenance (kernel vulnerability TCP Sack Panic). User’s impact: none.
Finished 18-06-2019 17:00 - 18:13 Maintenance scheduled to improve service backend resilience. Impact: UI/API will not be available. VMs are not affected.
Finished 04-06-2019 / 05-06-2019 Scheduled security maintenance downtime (MDS vulnerabilities)
Finished 02-04-2019 07:00-12:00 System maintenances & redundancy tests of SURFsara network components. No impact for running VMs.
Finished 08-03-2019 12:00 - 13:00 Unscheduled maintenance. Performance degradation of the hardware hosting UI and 2FA service components. Requires a reboot and consequent unavailability of these services. No impact for running VMs.
Finished 12-02-2019 08:20-08:40 UI/API downtime for maintenance. VMs not affected.
Finished 11-12-2018 / 11-12-2018 Scheduled downtime for security maintenance (L1TF)
Finished 10-07-2018 / 11-07-2018 Scheduled downtime for security maintenance (Spectre/Meltdown)
Finished 02-03-2018 / 05-03-2018 Scheduled downtime for security maintenance (Spectre/Meltdown)
Finished 10-11-2017 12:36-12:56 Scheduler did not place virtual machines, running virtual machines not affected.
Finished 11-10-2017 10:30-10:45 The UI does not allow logins.
Finished 02-10-2017 / 06-10-2017 Scheduled maintenance. Major overhaul of the underlying network infrastructure.
Finished 14-08-2017 Solved: Outgoing SSH traffic was blocked since 26-07-2017.
Finished 26-07-2017 : 16:45 / 26-07-2017 : 21:15 Unscheduled maintenance. Due to unexpected problems with one of our switches, the UI and part of the VMs on the HPC Cloud were disconnected from the internet. No network traffic in or out was possible. Internal network traffic, between VM’s appear to have been unaffected. At 21.15 the problems were solved.
Finished 27-06-2017 / 27-06-2017 Scheduled maintenance; there were a few short interruptions of the UI. No influence on running VMs
Finished 21-06-2017 : 15:35 / 21-06-2017 : 16:49 Unscheduled maintenance. During this period we experienced several disk issues causing requests being blocked in our storage systems. At 16:45 we removed the disks causing this issue, the storage systems became stable at 16:49. Running VMs were not affected.
Finished 18-04-2017 / 19-04-2017 Scheduled maintenance
Finished 20-02-2017 : 14:35 / 20-02-2017 : 15:30 Unscheduled maintenance. At 14:35 we received alerts that one of our backends was failing due to memory errors. To reduce load on our environment we temporarily closed all logins. These services were moved to the other node at 15:00, after this was completed we checked the environment. We opened our HPC Cloud again for logins at 15:30. Running VM’s are not affected.
Finished 24-01-2017 / 25-01-2017 Scheduled maintenance
Finished 02-09-2016 / 09-09-2016 moved to new data center
Finished 26-07-2016 / 27-07-2016  
Finished 05-04-2016 / 06-04-2016  
Finished 02-02-2016 / 03-02-2016