(Guest Post)
This is a continuation of my experience testing, and now putting into production, integrating Veeam Backup & Replication with HPE Nimble Storage. While testing, I came across a few bugs (post here), but once I worked through them, I was able to go full steam ahead in implementing my storage into Veeam to meet my organizational Availability goals. My case study is below. Would be interested in hearing everyone else’s experience in the comments.
Organization: North Kansas City School District (NKCSD)
Location: Kansas City, MO
End User Size: Approximately 20,000 students and 5,000 employees
Environment: North Kansas City Schools is fully virtualized, utilizing the latest VMware vSphere build, version 6.5. North Kansas City Schools runs its vSphere virtualization environment on HPE Proliant DL 360 Gen 9 servers and HPE Proliant DL 380p Gen 8 servers. Virtual Machines (VMs) are stored among two HPE Nimble CS500 hybrid (adaptive flash) arrays at a Central Office (DC1) site and remote (DC2) site. Veeam Backup & Replication (Veeam), version 9.5U2, is used for the organization’s availability and recoverability requirements, utilizing HPE Nimble CS440G-X2 and CS240G-X2 hybrid arrays for backup repository storage.
Challenges: Though the organization has a valid Availability (backup and recovery) Plan, with recent changes to the software and hardware the organization currently utilizes (Veeam Backup & Replication & HPE Nimble Storage), the organization would like to further streamline Availability operations utilizing a central management interface for backup/replication processes, as well as simplify recoverability tasks if the need arises, using the Veeam Backup & Replication Console
Goals: Institute a local backup strategy that incorporates organizational Service Level Agreements (SLAs) based on Recovery Point Objectives (RPOs); institute a backup strategy to have multiple copies of data to mitigate a Single Point of Failure (SPOF) scenario, as well as transfers production data to a remote site; institute a disaster recovery (DR) strategy for the ability to retain business continuity in the event of a primary datacenter disruption or disaster
Summary: Content below discusses how NKCSD utilized both Veeam and HPE Nimble Storage to overcome Availability challenges and meet their Availability Solution goals
PREREQUISITES & ASSUMPTIONS
1. Veeam Backup & Replication is installed
2. Physical Windows Servers are in place to be used by Veeam as both Backup Proxies and Repositories. To be able to integrate HPE Nimble Storage with Veeam, these servers have Microsoft’s iSCSI Initiator installed to see the source and target Nimble storage. Repository Transport Mode used is Direct SAN
3. Nimble arrays are installed and configured with iSCSI Initiator Groups of the Veeam Backup Proxy servers, assigned to the Volumes housing the VM data wanting to back up, as well as Volumes used for backup Repositories. Replication partners were configured & tested as well
4. Full understanding how to use & configure HPE Nimble and Veeam Availability processes, such as – Nimble: storage Snapshots & Replication; Veeam: Proxy/Repository and Backup & Replication configurations & settings
Goal Resolution
1. To meet RPO SLA and Retention goals, NKCSD created regular Backup Jobs and configured the Restore Points (RPs) accordingly based on the permitted backup schedule (between 7a-7p, M-F; 8p on Sa and Su). If longer retention was needed for certain applications (VMs), a Backup Copy Job was created and configured according to retention goals. Where only certain pieces of data within an application was needed for longer retention, a Backup Job was created to include only those files or folders within a VM to reduce storage waste and make more storage available for other backup or retention goals
2. To meet both its RPO goal, as well as multiple data copies/location goal, NKCSD configured these regular Backup Jobs to do a backup from Primary Snapshot to a second local array, as well as institute both HPE Nimble Storage Snapshot and Storage Replication. Implementing these features provides NKCSD a copy of its data on the source array, a full backup copy on a second local array, as well as a full copy on a remote array
3. Finally, to meet two goals of having a full remote copy, as well as provide an intuitive DR solution, whether a small DC1 outage or site-wide disaster, NKCSD implemented Veeam Replication. To help the organization achieve its Recovery Time Objective (RTO) goals, a second site (DC2) was created with appropriate software and hardware, as well as implemented a second Veeam server, used only for Replication jobs. Implementing this configuration assures intuitive recovery by having a central & familiar recovery management console, as well as requiring a minimal amount of recovery tasks needing to be performed to get applications back online to meet its RTOs
Considerations: Below are some challenges or considerations to take into account when configuring Veeam with storage integration into your environment. Some items were already mentioned above, but are worth repeating:
1. Backup Job configuration – test different Job Modes (Forever Forward Incremental vs Reverse Incremental) & Repository configurations (e.g. per-VM and/or align disk) to assure you are able to a) meet your RPO SLA goals, and b) to achieve the best Backup Job performance (processing rate)
2. Storage Snapshot and Replication – verify your production/source arrays have enough storage for snapshot data, and your remote arrays for replication data. When creating Volume Collection (VC) schedules on the HPE Nimble Storage arrays, make sure to configure the VC snapshot schedule far out (once a week) so as not to waste storage
3. Repository storage – verify you have enough target storage for data; this of course goes without saying, but keep this in mind especially if implementing the “per-VM” Repository Advanced Setting in Veeam as that configuration tends to use up more storage. If using a supported array with deduplication capability, this shouldn’t be much of a concern
4. Data – keep in mind you may not need to have long retention of all data (VMs). To best utilize your storage, have longer retention jobs only for most critical applications and/or applications required by law to have certain retention periods. To make better use of your storage even further, you can even configure Veeam “inclusion” jobs to backup only certain file(s) or folder(s) rather than waste space backing up whole VM data when its not needed
5. Data location – if you are at an organization with really aggressive RPO SLAs, and the only way to achieve those SLAs is to use storage snapshot orchestration jobs, keep in mind that is a SPOF as you have no complete copy of data, and no complete copy of data located on a second array or media type
Conclusion: With recent technological advancements of both Veeam Backup & Replication and HPE Nimble Storage, NKCSD was able to utilize Veeam integrated with Nimble to architect an Availability Solution to meet its Availability goals. Storage integration allows NKCSD to:
1. Create backup jobs to meet organizational RPO SLAs
2. Have multiple copies of data across multiple media types, configurable within those same backup jobs (i.e. no multiple backup jobs needed)
3. Have a satisfactory data retention solution utilizing Veeam Backup Copy or Veeam ‘Data Inclusion’ jobs
4. Have an intuitive, centrally managed DR solution in the event of a DC1 disruption or complete outage
5. Ability to configure and manage both backup & storage Availability tasks within one management interface – Veeam Backup & Replication Management Console
You can find & ping me on Twitter:
Shane Williford, @coolsport00
Sr. Systems Engineer