While discussing storage and data protection with some colleagues recently I asked the question “At what point does your storage system become too big to fail?” Prior to this, we had discussed various ways to achieve aggressive RPO/RTO for varying data sets. Between features and rich backup products like Veeam, Unitrends and intelligent storage snapshots being baked into appliances like Nimble Storage, you can architect a plan that will offer a plethora of recovery points and recover the data with near immediate access with little manual effort. Add into the mix one of the more exciting features from VMware, Virtual Volumes (VVOLs), and you start to see how these options begin to complement each other. When used in conjunction you can start to define objectives on a very granular level, which tends to shrink the amount of time it takes to gather the data. This in turn has a positive effect on the impact backups have on your environment: less data = less traffic = less contention.
This is great news for smaller sets of data, but year over year, our data footprints are growing tremendously. So data grows and grows, but the backup window that admins have to work with stays the same. At some point we may need to adjust our backup jobs to be able to protect the data that’s critical to our business to abide by these constraints. The easy to create, all-encompassing jobs that were setup a year ago might start to take too long to complete as data grows. These will get adjusted so that we’re backing up just the files, rather than the entire VM or physical server without impacting production. Often, this will have an adverse effect on the RTO, and if not that, then certainly the ease of recovery. You may no longer be able to recover an entire SQL Virtual Machine with a single click, just the database it runs. As a result of this, recovery testing (you’re doing this, right?!) takes longer as well.
So what’s an Admin to do? Setting up file or mailbox size quotas won’t gain you any friends. Selective jobs certainly gain you time during the backup, but increase the time for recovery. Moving the data faster might work, but for how long and at what cost? There are Public Cloud options but they have their own list of considerations to come to terms with.

Great Article Artificial Intelligence Projects
ReplyDeleteProject Center in Chennai
JavaScript Training in Chennai
JavaScript Training in Chennai