Since Veeam B&R v9, there is a support of HPE StoreOnce Catalyst storage. HPE StoreOnce Catalyst integration improves the time it takes to boot a VM from backup file residing on StoreOnce about two times comparing to Instant VM Recovery from backup files on a StoreOnce NAS Share. This post is about using a Catalyst store as a replication source, Veeam ‘s documentation does not mention it in scenarios. Basically, it does not explain the automatic election algorithm of a Gateway server during a replication job (Veeam Help Center).
Virtualization infrastructure overview:
Infrastructure relying on 3 physical sites:
- Two Datacenter 1km far from each other running a VMware stretched cluster on HPE 3PAR, this is the logical primary Datacenter (production).
- A Test/Dev and DRP Datacenter 150Km far away from the two other holding the same stack as one of the primary sites. It also has an HP StoreOnce array.
- 1 x 10Gb/s between each primary site and the DRP site.
Veeam production infrastructure overview:
- 1 physical dedicated server per site (primary and DRP) doing Direct SAN Proxy, Repository, Interaction proxy and Mount Server.
- 1 virtual Veeam server running on the DRP site.
- 1 unique HPE Catalyst store mapped to two Veeam Repository objects to maintain a proxy/site/repository affinity.nning on the DRP site.
- 1 unique Catalyst store mapped to two Veeam Repository objects to mantain a proxy/site/repository affinity.
Backup process is :
- First, backup from HPE 3PAR to local site repository using Direct SAN and Storage Snapshots integration. Hold 35 days on site.
- Then, at the end of the backup window, starting a Backup Copy job on each site copying the freshly done backup to an HPE Catalyst store at the DRP site (with a complete retention schedule over two years).
- Finally, at the end of the Backup Copy Job, rebuilding the entire production systems on the DRP site with replication jobs using the Catalyst store. Using no resources from production sites.
Everything just works as expected (lightening fast) and in the case of a major outage on both production site, the end-user would be able to restart production systems at the DRP site in a very short time. It also allows them to validate Backups and DRP in one single operation because theirs SureReplica jobs are based on replicas that are based on backups which mean you validate the whole chain./p>
The “thing” is about the Gateway Server, documentation has no information about the election process when used in replication job. Here is Eric’s (Veeam Vanguard) understanding based his feedback:/p>
- You have to let the gateway server be elected automatically because you can’t have multiple preferred gateway servers per Catalyst repository.
- You have to make a Proxy affinity between HPE Catalyst repository and the DRP site proxy.
- Replication job from a Catalyst store uses the target proxy as the gateway server when setting in automatic selection.
Thanks to @Eric for this useful story. 🙂
You can learn more here.
Veeam Agent for Windows step by step installation guide.
If you need to disable UAC Server on Windows Server 2016 Script Guide.
If you need to upgrade your Veeam Backup Server (v9 or v9.5) you can follow this article.