Veeam storage functions
Is Veeam the most significant storage company that no one knows about?
One of the essential features of any Veeam Backup and Replication architecture is the storage hardware that the backup target or backup repository will sit on. The repository performance, features, and scalability are critical to ensuring that the Veeam deployment will meet customers' needs for quite a while.
Unfortunately, this is the component that many customers tend to make bad decisions, usually due to VARs and storage vendors trying to make the backup decisions for the customer. This is easy for the partner and storage vendor community to do because, as we all know - we at Veeam pride ourselves on the fact that the Veeam Backup and Replication suite is a software-defined and hardware agnostic backup storage platform. The partners, storage vendors, and even customers don't realize that Veeam also has another tag ling used internally - "Veeam is the largest storage company that no one knows about"! I first heard this a few months ago during an internal training seminar, then at an internal training at our main R&D office in Prague. Naturally, I had to investigate this further and figure out how I could help add this to my architecture and educate our partners, alliance vendors, and customers.
Below is a list of features that can be used with high disk dense servers (Cisco s3260, HPE Apollo 4510, DELL EMC R740 XD, Lenovo SR650 V2, etc.) and entry-level and dumb storage arrays to make a feature-rich performant Veeam backup target. That means less money spent on the storage and less spent overall on a Veeam solution.
Veeam storage features
Scale Out Backup Repository (SOBR)
This is one of my favorite features. The SOBR allows Veeam users to take multiple storage units and add them to a single namespace. This allows Veeam backup jobs to target a "pool" of storage, and Veeam will handle the data placement to the proper extent.
What I like about using a SOBR is that if you use a dense storage server like a Cisco S3260, HPE Apollo, DellEMC r740XD, or Lenovo Sr650 V2 is that you are adding, network, compute, and storage to the pool, so the backups never slow down as you scale the backup target pool. Oh ... and the best part - these boxes have a large amount of storage (think 700 TB +) for an excellent price.
SOBR can also provide the ability for NAS backups, RMAN, and SAP Hana backups to spread the backup streams across multiple storage extents to improve backup speed performance.
Capacity tier is what Veeam calls the automatic tiering of backup files to object storage. This supports a copy, move, and copy and move feature. Also, the capacity tier allows for making backups copies storage on S3/S3 compatible storage that supports object lock as immutable.
Veeam has had this feature out for a few years and a few versions already. Many arrays support auto-tiering on the market today but are not supported by Veeam, as Veeam needs to do all the tiering to ensure proper files and blocks are moved and tracked correctly.
Veeam can now take Grandfather, father, son (GFS) point backups from the capacity tier and move them into long-term, low-cost cloud archive storage like AWS Glacier and Azure Deep Archive. This is an automatic process through the Veeam SOBR.
Veeam uses multiple data reduction strategies:
- Skipping windows swapfile blocks
- Skipping Windows dirty blocks
- In-line dedupe and compression - reduces the amount of data that gets written over the network to the repository and minimizes the amount of data that needs to be ingested by the backup target. This can speed up backup data transfer from the proxy to the storage.
- REFS/XFS fast clone/Block cloning - Fast Clone is the Veeam Backup & Replication technology that helps create quick file copies. Fast Clone increases the speed of synthetic backup creation and transformation, reduces disk space requirements, and decreases the load on storage devices. Fast clone helps to reduce storage consumption and IOP overhead during synthetic operations like retention point merges and artificial full backup files.
Veeam has several features to help with storage performance. This includes making writes and reads faster, and it helps with the throttling of the backup speed to support both the backup target and the source storage.
- All Veeam data mover transport modes support async IO
- Veeam data movers send data directly to the storage systems RAID controller onboard battery-backed cache, which allows for faster writes
- Async IO writes
- The Veeam hot add proxy and backup from SAN snapshot backups also have a read-ahead block algorithm and storage queue depth for a continuous data stream transfer to the repository.
- Veeam Storage Latency Control adjusts the number of threads and backup transfer speed to match the set latency threshold observed on the production datastores.
- Veeam Storage level corruption guard/health check - An automatic health check can help you avoid a situation where a restore point gets corrupted, making all dependent restore points corrupted. If during the health check Veeam Backup & Replication detects corrupted data blocks in the latest restore point in the backup chain (or, in case of forever forward incremental and forward incremental chains, the restore point before the latest one if the newest restore point is incomplete), it will start the health check retry and transport valid data blocks from the source datastore to the backup repository. Depending on the data corruption scenario, the transported data blocks are stored to a new backup file or the latest backup file in the backup chain.
- A hardened repository provides a way of hardening a Linux-based repository(s). This is done by removing the need to use root, su, or sudo during backups. It also eliminates the requirement to have SSH/SCP access available. When Veeam configures the repository, it uses a single-use credential to access the Linux repository and configure the Veeam components. The Linux repository can be a standard repository or be some extent in a SOBR.
Since Veeam provides all these storage functions, you should consider using a storage dense server, DAS/JBOD servers, as your next Veeam repository.
Thanks for sharing this Joe.
Too true, I get a lot of customers talking about backing up “directly to the cloud”, as they’re worried about CapEx costs for hardware, power, cooling & colo costs and forklift upgrade approaches.
Everytime I explain to them that they’re omitting backups key task: restoration and the associated RPO & RTOs which get throttled via WAN.
Dense Tier 1 backup storage is still the best solution. Where customers can afford I suggest a higher IO backup tier for short term backups, if no desire for replication or CDP, then a separate SOBR for traditional dense storage spinny that has longer retention, then let them leverage capacity/archive on that.
As harsh as it sounds, I want Veeam to remain the storage vendor nobody knows about, as if Veeam got into the hardware appliance space, it’d slow down their innovation IMO!
I agree with you on the hardware vendor space
@MicoolPaul! I do think that Veeam can continue to innovate in the storage related features that get added to the product to help with backup repository performance and features.
Isn’t this piece just informative? Yes it is… Thank you so much for sharing this educative piece with us
Great article and remember
@vmJoe , thanks!
I’ve said this another way in that Veeam is at the “crossroads of data”; maybe we’re the traffic light. LOL