Upcoming Maintenance window at UC

Resolved Posted by Michael Sherman on May 25, 2022
Outage start Monday, June 06, 2022 8 a.m.
Expected end Tuesday, June 07, 2022 6 p.m.

Update 5:30 pm: Issues are resolved, all nodes are usable again.


Update 4pm June 7th: Provisioning of baremetal nodes is restored. We're seeing failures to create leases for P2 nodes (types compute_skylake, gpu_rtx_6000), but reservation of P3 nodes is succeeding.


Update: 7pm June 6th: The upgrade has been completed, and existing instances, Jupyter, Trovi, and the docker registry are back online. Provisioning of new instances is still unreliable, and work is in progress on fixing that.

Since only new instances are affected, we're downgrading the outage severity.


On June 6th, Chameleon services at UC will be unavailable to permit an OpenStack version upgrade.

The CHI@UC horizon webui, openstack api, and all running instances will be unavailable for the duration. In addtion, as Trovi, Jupyterhub, and the Chameleon docker registry will be unavailable, as they depend on the CHI@UC object store. These services should be restored sooner than the main CHI@UC site, but may be unreliable for the duration of the upgrade.

We're taking steps to mitigate the disruption, and the above list may shrink as we approach the maintenance window. We will update this announcement accordingly.