What are the key challenges companies face during a Backup upgrade or refresh? Evros Solutions Architect Stephen Seagrave explores the most common pitfalls with today’s Backup solutions:
Backup technology has evolved dramatically over the past decade, moving on from the traditional legacy solutions that copied to tape on a daily basis. Today, backup is mostly directed to disk, and we infrequently use tape if at all. Backup solutions now deliver capabilities such as automated testing, instant restore and labs for testing application updates, patching and various restore scenarios. Solutions such as Veeam DataLabs with SureReplica and Sure Backup are driving a new availability infrastructure that embraces Disaster Recovery with instant availability and many more resource intensive features. The result is that your backup solution has now moved from a back-end protection only product, to a front-of-house production availability and lab environment service. To fully appreciate and exploit the capabilities of modern-day backup solutions, you must consider it as a production enhancing service, with features your environment never had before. My job is to make sure your company is making the most out of it all, and to avoid the common pitfalls that many companies face when tackling a Backup upgrade or refresh.
The 5 Most Common Pitfalls in Backup:
Under-investing or Underestimating
Often, companies under-invest or underestimate the requirements in hardware for the backup solution, whether backup storage is in disk, cloud or tape. I see companies invest in a back-up software solution for the headline features such as Veeam data labs and instant restore. Yet, they’re unable to implement these features fully because they don’t have the infrastructure to utilize it. If you expect to run an application stack of VMs from your backup files, your backup infrastructure must have the power required.
Using Old Hardware
If you stumbled on the first pitfall, then the second scenario is soon to follow. This time, you’re now expecting far too much of the infrastructure you have. Often, we’re seeing organizations with the backup solution sitting upon repurposed old infrastructure. It’s not uncommon. Yet, if you re-use old infrastructure such as an old ESX or Hyper-V host for backup servers, you may not have the capability to deliver your expectation, which leads to disappointment and then imminent failure. Which then leads to reinvestment and the extra cost of redeployment days to complete what should have been accomplished in the first place.
Mismatch of New Backup Solution vs Old Production Hardware
Expecting your new backup infrastructure to achieve more than your production infrastructure is capable of, is a common complaint. You may have invested in a backup server which can achieve exactly what you want, such as a 5-minute backup schedule. Yet the production SAN is old and under spec’d and cannot give the IOPs (In, Out Processes) required. The result: Read speed from the old SAN is too low and backup has to be dialed back, so your production VMs are not impacted by the backup speed. This can also happen when customers are getting rid of tape and move to a cloud backup solution but didn’t take into account the internet bandwidth required. Customers end up with an all-singing, all-dancing backup server that can’t draw from the source, or push to the destination.
Know your Compliance Requirements
Another common pitfall is misunderstanding the capabilities of what’s possible and choosing an incompatible solution because of lack of clarity around the business requirements and compliance guidelines. Often, companies’ compliance requirements are not called out in advance which leads to incurring costs such as extra labor days or more hardware kit needed. Make sure you clarify all requirements and don’t assume they will implement a certain feature in your deployment of the backup software. Differing features may require specific setup and may also have software or hardware dependencies.
For example: Veeam’s SQL log shipping and Point-in-Time restore allows you to restore a database to a specific second of time, or even a specific transaction. To do this, Veeam needs to mount the full database backup into an SQL mount server and replay the transaction logs to get to that point in time. As the default Veeam install is on SQL express (maximum database size), it can mount to 10GB, so if your production databases are larger, you will need to have a full version of SQL deployed to use as a mount server. Finding this out after the fact will add cost and time to a project.
Verify Features with your Solution Provider
Each Backup solution will have very similar features but are implemented very differently. Make sure you’re choosing the right provider who can implement the correct software solution in a process that makes sense, and won’t be over-cumbersome to your current work operations. Arcserve has an Instant Restore feature and so does Veeam. In Arcserve, this is more like a replica on standby but uses full production space on the SAN. To get instant restore for all your VMs, this would mean you have to double the size of your production storage. Veeam’s Instant Restore runs from the backup files mounted as a datastore so your backup server needs to be fast. Both software solutions work and both boast the same capabilities on marketing material, but they work very differently and require vastly separate hardware to function. Work with a trusted partner and your architect to go through how each software solutions does each of those headline features. Then prioritize the features that matter to you and your business and pick the backup solution that suits your company best.
One thing you can do to avoid all these pitfalls is to engage with a Solution Architect specialized in Backup and Availability solution technologies. The specialist will define all the ups and downs, ins and outs of each solution, allowing you to implement the best solution for your company, and for your budget.