- 24-04-24Status.bit.nl in nieuw jasje!
- 12-04-24Nieuw bij BIT: GPU hosting
- 25-03-24BIT breidt netwerkconnectiviteit uit met aansluiting op NL-ix^2
- 13-03-24No More Leaks: Samenwerken tegen cybercriminaliteit
- 03-03-24Geen verandering twee jaar na invoering sancties tegen Russische media: FOIC roept (wederom) op tot einde van ondemocratische censuur
- 29-01-24Onzichtbare upgrades
- 16-01-24BIT's Wido Potters wint Felipe Rodriquez Award voor inzet voor privacy
- 10-01-24BIT en partners zetten de koers voor een gedecentraliseerde Europese cloud met ECOFED-project
- 02-01-24Onze eigen stroom inkopen: de resultaten van 2023
- 24-11-23BIT Friday 2023
Maintenance on spine and storage switches
29-09-2022 00:00:00
- 29-09-2022 06:00:00
Urgency: Planned
Affected services:
Reference number: 181560
Summary:
The spine switch in BIT-1 and the switches of the storage clusters in BIT-1 and BIT-2C will be fitten with new firmware. Traffic via these switches will be rerouted prior to the maintenance. iSCSI LUNs and CephFS volumes can experience short delays several times.
Details:
All switches in BIT's data center network are connected to two spine switches. During the maintenance, new firmware will be installed on the spine switch in BIT-1. As a result, the switch will be unavailable for at least 15 minutes. Traffic is automatically rerouted via the switch in BIT-2A.
The storage switches in BIT-1 and BIT-2C will be provided with new firmware in turn. Prior to the upgrade of a switch, the iSCSI-LUNs on the storage gateways will be switched over to the other two storage gateways. During a switch-over the iSCSI targets can be unreachable for a few seconds, causing iSCSI-LUNs to respond with a short delay. MS SQL instances and MS SQL databases use iSCSI storage. These services may therefore be briefly experienced as unavailable due to delayed responses. This maintenance will also have an impact on CephFS volumes. The metadata servers are directly patched to the storage switches. Prior to the upgrade of the relevant switches, the standby metadata server, which is not patched to the switch being updated, will be made active. This can cause a momentary delayed response of the volume. After the switchover, IO will resume as normal. At the time of maintenance, no standby metadata server will be available and the setup is therefore less redundant at that time.
Update placed at 21/09/2021, 2:45h
This maintenance was completed succesfully at 2.00h.
Affected services:
- Internet Access
- Virtual machines
- Shared storage
- Shared hosting
- iSCSI LUNs
- MS SQL instances
- MS SQL databases
- Reduced redundancy
- iSCSI LUNs might not respond for a short moment
- MS SQL instances might not respond for a short moment
- MS SQL databases might not respond for a short moment
- Shared filesystem (CephFS) might not respond for a short moment
Reference number: 181560
Summary:
The spine switch in BIT-1 and the switches of the storage clusters in BIT-1 and BIT-2C will be fitten with new firmware. Traffic via these switches will be rerouted prior to the maintenance. iSCSI LUNs and CephFS volumes can experience short delays several times.
Details:
All switches in BIT's data center network are connected to two spine switches. During the maintenance, new firmware will be installed on the spine switch in BIT-1. As a result, the switch will be unavailable for at least 15 minutes. Traffic is automatically rerouted via the switch in BIT-2A.
The storage switches in BIT-1 and BIT-2C will be provided with new firmware in turn. Prior to the upgrade of a switch, the iSCSI-LUNs on the storage gateways will be switched over to the other two storage gateways. During a switch-over the iSCSI targets can be unreachable for a few seconds, causing iSCSI-LUNs to respond with a short delay. MS SQL instances and MS SQL databases use iSCSI storage. These services may therefore be briefly experienced as unavailable due to delayed responses. This maintenance will also have an impact on CephFS volumes. The metadata servers are directly patched to the storage switches. Prior to the upgrade of the relevant switches, the standby metadata server, which is not patched to the switch being updated, will be made active. This can cause a momentary delayed response of the volume. After the switchover, IO will resume as normal. At the time of maintenance, no standby metadata server will be available and the setup is therefore less redundant at that time.
Update placed at 21/09/2021, 2:45h
This maintenance was completed succesfully at 2.00h.