Group: Forum Members
Posts: 8.9K,
Visits: 59K
|
Having just written a few posts over the last few days about GFS, FS (with and without "variable granularity" and Synthetic Fulls), and Incs Forever, I too am starting to realize just how nuanced the choice of backup strategy can be, and therefore how confusing it can be to the average Joe. The fact is that Reflect offers a high degree of both capability and flexibility, and that combination almost always leads to a certain degree of complexity.
But in terms of "promoting" Incs Forever, although I use that strategy myself because I appreciate a variety of its benefits, I'm not sure it's most likely to be best for the average user. I'm assuming that the average user will not be using a disk rotation like I do and therefore will want to retain more backups on their destination than I do on any one of the 9 disks in my rotation. So if the user wishes to perform daily backups and retain 3 months' worth, that would be ~90 Incrementals. I know that there's nothing technically wrong with that, but a chain that long would make me feel uneasy because it increases both the risk and impact of "bad data" somewhere in the chain. I also wonder how it would affect restore/browse times, especially now that Delta Incremental Indexes is enabled by default. But far more importantly, in this scenario, all backups are part of a single chain, which increases risk on its own but becomes an even greater risk when the chain is long. To be fair, I've been using Incs Forever with Synthetic Fulls for 10 months of daily backups now and have not had a single failed consolidation operation, but that's not the only possible point of failure. Hard drives can develop bad/damaged sectors due to manufacturing flaws or "external factors", e.g. the user dropping the drive or (worse) knocking it a desk while it's running if the drive doesn't have a freefall sensor built in. With Incs Forever and only a single destination, this could render a huge portion of the user's backup history useless, up to and including the entire chain, whereas a GFS strategy might limit that damage to a month's worth of backups.
And from a pragmatic standpoint, the average user just wants their data backed up. They're probably less concerned about: - How long backups take (they can run overnight and/or at lower priority if needed) - How much storage backups consume (storage is cheap)
They may not even be concerned about how long backups are retained, since in most cases people want to browse/restore recent backups, not ancient backups. And even if retention is a significant factor, the retention policies are fairly simple to understand in my opinion and are also somewhat independent of the backup strategy selection. Lastly, the average user probably does not understand, nor should they have to understand (in my opinion) the difference between Full, Diff, and Inc backups. Rather, they pretty much just want to know what points in time they have available from which to restore data, and exactly how that is achieved is unimportant to them.
Given that, pretty much ANY template the average user chooses is going to get them where they want to be. It very likely will NOT be the "best" strategy for their purposes in terms of time to perform backups, storage consumed, risk/impact of bad data, etc., but if that starts mattering to them, then as you say, Reflect already describes these strategies in the UI and has excellent documentation elsewhere that describes them in greater depth in order to help those users make informed decisions. It's not reasonable for people to expect they'll be able to utilize powerful tools to their fullest extent without educating themselves a bit, after all.
|