Folks , I am exploring the possibility of OpenStack to set up an On Prem DC and Near line DR using OpenStack . I have exposure to OpenStack from the Ussuri version onwards but am not a master of OpenStack. Objective is to achieve near zero data loss of the growing VMs at the DC replicated in real time to the Near Line DR. At times of a DC crash or for scheduled maintenance activity to ensure business continuity (RPO near zero and RTO 30 minutes - 1 Hour) to operate from the DR site. Heard about similar solutions exists for VMWare or any proprietary OEM like vSphere replication,vSAN and SRM to orchestrate failover and fallback with RPO near zero and RTO 30 minutes to 1 Hr. But not an option due to budget constraints at the moment. What ways do folks at OpenStack achieve this kind of requirement? The underlying tools and techniques I have chosen to deploy and the best possible design of DC and DR to achieve the objectives. Any hints and guidance to start, much appreciated. Thank you, Krishane
From what I've seen Openstack doesn't have a good way of handling this, Kind of like running in AWS or another cloud provider, treat your VMs as throwaways at any moment and run copies. You can do cross DC clusters with Ceph but thats only for storage, not the compute site of things. [cid:c76a487b-1e23-4a00-95f0-56e81ec63437]<https://outlook.office.com/bookwithme/user/661ab6f82f554e50a92cefcba55614b6@rit.edu?anonymous&ep=bwmEmailSignature> Book time to meet with me<https://outlook.office.com/bookwithme/user/661ab6f82f554e50a92cefcba55614b6@rit.edu?anonymous&ep=bwmEmailSignature> ________________________________ From: KK CHN <kkchn.in@gmail.com> Sent: Wednesday, April 23, 2025 3:27 AM To: openstack-discuss@lists.openstack.org <openstack-discuss@lists.openstack.org> Subject: Deploying a DC - DR solution Folks , I am exploring the possibility of OpenStack to set up an On Prem DC and Near line DR using OpenStack . I have exposure to OpenStack from the Ussuri version onwards but am not a master of OpenStack. Objective is to achieve near zero data loss of the growing VMs at the DC replicated in real time to the Near Line DR. At times of a DC crash or for scheduled maintenance activity to ensure business continuity (RPO near zero and RTO 30 minutes - 1 Hour) to operate from the DR site. Heard about similar solutions exists for VMWare or any proprietary OEM like vSphere replication,vSAN and SRM to orchestrate failover and fallback with RPO near zero and RTO 30 minutes to 1 Hr. But not an option due to budget constraints at the moment. What ways do folks at OpenStack achieve this kind of requirement? The underlying tools and techniques I have chosen to deploy and the best possible design of DC and DR to achieve the objectives. Any hints and guidance to start, much appreciated. Thank you, Krishane
At CERN, we run OpenStack and ceph across two data centres to address some of our needs for resilience. Some background on the approach is at https://inspirehep.net/files/18ee7c0191e1d04bc99e068c44967b49 and https://indico.cern.ch/event/1338689/contributions/6010769/ (presentation and paper at the bottom). We’ve set up two OpenStack independent regions and have investigated various approaches such as rbd mirroring and multi-site S3 to provide redundancy. We continue working with upstream in areas such as CephFS replication/snapshotting. It may be worth getting in touch with the OpenStack VMWare migration working group (https://www.openstack.org/vmware-migration-to-openstack/) to see how others are investigating achieving similar goals. Tim On 23 Apr 2025, at 09:27, KK CHN <kkchn.in@gmail.com> wrote: Folks , I am exploring the possibility of OpenStack to set up an On Prem DC and Near line DR using OpenStack . I have exposure to OpenStack from the Ussuri version onwards but am not a master of OpenStack. Objective is to achieve near zero data loss of the growing VMs at the DC replicated in real time to the Near Line DR. At times of a DC crash or for scheduled maintenance activity to ensure business continuity (RPO near zero and RTO 30 minutes - 1 Hour) to operate from the DR site. Heard about similar solutions exists for VMWare or any proprietary OEM like vSphere replication,vSAN and SRM to orchestrate failover and fallback with RPO near zero and RTO 30 minutes to 1 Hr. But not an option due to budget constraints at the moment. What ways do folks at OpenStack achieve this kind of requirement? The underlying tools and techniques I have chosen to deploy and the best possible design of DC and DR to achieve the objectives. Any hints and guidance to start, much appreciated. Thank you, Krishane
participants (3)
-
Forrest Fuqua
-
KK CHN
-
Tim Bell