Advice: NeCTAR-wide outage for Neutron: 2015-10-06 (08:00) to 2015-10-08 (08:00)

Follow
The following announcement has been sent to all NeCTAR users:
Dear NeCTAR Research Cloud User,
This email is to inform you of a scheduled outage to parts of the NeCTAR Research Cloud service. The outage will allow us to perform a cloud-wide upgrade that is the first step towards providing advanced networking services (eg. floating IPs, private networks and VPNs). These improvements are a result of user feedback and extensive planning has occurred to minimise the impact.
DURATION: 2 days 0 hours
START TIME: 2015-10-06 08:00:00 AEST
END TIME: 2015-10-08 08:00:00 AEST
Updates will be posted to the support news page here.
IMPACT:
The following impact will be experienced across all Research Cloud zones during the outage window:
  • All active instances will continue to run
  • There will be a small 1-2 second outage of all running instance's network connections
  • There will be some short outages of the instance metadata service (the cloud internal http://169.254.169.254/)
  • The Research Cloud control plane (web dashboard and APIs) will be disabled for users during the outage window, this means no cloud control actions can be performed on existing instances (e.g. terminate, snapshot) and no new instances can be booted.
It is anticipated that the majority of active instances will not be affected by the 1-2 second network outage (e.g. this will not interrupt TCP connections).
Extensive tests indicate that the control plane (web dashboard and APIs) will be unavailable for only part of the first day of the outage window. An additional day has been allowed for unexpected issues and roll back, if required.
All NeCTAR staff will be on standby during and immediately after the outage. If users notice any issues with existing instances they should report them to NeCTAR Research Cloud support.
Your current instances are listed below.
<Deleted>
ACTION REQUIRED:
Although no to minimal impact is expected to user instances, it is recommended that for critical services users should complete their own backups or snapshots, or otherwise copy critical data off the Research Cloud.
Users are advised to complete necessary management tasks via the APIs or the Dashboard before the outage window. Automated scripts that create and manage instances via the APIs will not operate during the control plane outage.
Users may wish to temporarily set passwords on selected local accounts in their instances. This will enable console login and recovery without reboot in the event that the instance loses networking (this may happen e.g. due to a DHCP client bug inside the instance).
Users are also advised that the Swift object storage service will not be affected and remain available during the outage window.
IMPROVEMENTS SCHEDULE:
NeCTAR Core Services is pleased to announce the staged introduction of a new network service for the Research Cloud called Neutron. Full details on the OpenStack Neutron project can be found here: https://wiki.openstack.org/wiki/Neutron.
In preparation for this release NeCTAR Core Services will also be upgrading the cloud software to the latest OpenStack release called Kilo. Further information on the benefits of this upgrade can be found here: https://wiki.openstack.org/wiki/ReleaseNotes/Kilo.
These upgrades are complex and have a number of dependencies, hence Core Services is taking a phased rollout approach over a number of stages as follows:
Stage Description Release Date
Kilo Upgrade Upgrade of OpenStack Nova Service in preparation for for the introduction of the Neutron Service (There is no outage or impact to users for this upgrade) 23 Sep - 5 Oct
Neutron Stage 1 Preparation for networking services to be deployed, introduce global security groups (This is the only stage requiring an outage) 6-7 Oct
Neutron Stage 2 Basic network services available such as private tenant networks between instances, virtual routers and floating public IP addresses Q4'15 - Q1'16
Neutron Stage 3 Advanced network services such as load balancers, firewalls, VPNs, etc Q1'16 - Q3'16
 
SUPPORT:
 If you have any queries regarding this outage, please contact us:
support@nectar.org.au
1300 080 431

We apologise sincerely for any inconvenience caused by this outage and appreciate your patience as we continue to improve and evolve the NeCTAR Research Cloud for the benefit of Australian researchers.

Regards,

The NeCTAR Research Cloud Support Team
--
This email has been sent to users of the NeCTAR Research Cloud. It has been sent to the email address associated with your Research Cloud login. These emails are essential communications which we endeavour to keep to a minimum. They only relate to events that may affect your NeCTAR Research Cloud resources or data integrity. If you receive multiple copies of this email, it may be due to the fact that you are a member of multiple projects.
Have more questions? Submit a request

Comments

Powered by Zendesk