OneDrive Sync Mechanics
-
I just don't see the argument for not wanting to have another copy of the data in another place.
I acquiesced to your concept of not having data locally. I don't personally like it because I think it makes working on stuff (using older or non-cloud aware apps) harder. (Yes, that is another thing I'll need to acquiesce on.)
But I don't understand why it wouldn't make sense to have another copy of my entire OneDrive file structure in another cloud service, like Amazon Cloud Drive.
-
@BRRABill said:
Hence the need for backup.
You can use a complicated backup mechanism for this, to handle backup versions. But this is overkill and, to some degree, archaic. The backup system doesn't understand your triggers, applications do. Versioning should be done closer to the data and backups be at a high level for the most efficiency and functionality.
Pictures are a difficult one because they are hard to deal with because you don't version them like that. Having a read only system is the best bet, normally.
-
@BRRABill said:
I just don't see the argument for not wanting to have another copy of the data in another place.
Ah, you are getting my point wrong. I never said that at all. I'm saying that you always need a copy in another place. I'm saying that you don't need to be the one that takes that copy.
-
@BRRABill said:
But I don't understand why it wouldn't make sense to have another copy of my entire OneDrive file structure in another cloud service, like Amazon Cloud Drive.
Again, would you tell your users not to trust you? If not, why do you feel that you should act differently than you advice others to act?
-
I am not arguing that what you are saying doesn't make the most sense.
But haven't you ever had something go wrong when it was good to have TWO backups of it?
I'll give you two examples.
Back in the day of tapes, I once had a backup job running that was looking for a tape. It had jumped a day, and long story short I put a tape in to do a restore, and the data got wiped out. Since then, I used the write tab every time I did a restore. Did I NEED to do that? No, but since I had been burned, it made for one more level of protection.
Also, back when iPhones were backed up to the machine, I had many instances where it would do an update, then backup the bad data. After getting burned a few times, I learned to do a backup of the iPhone, then do a backup of that backup. Overkill, perhaps. But I got burned enough times for it to make sense.
I am sure we've all been burned at one time or another by our friendly computers. Having a backup of the backup I think makes absolute sense.
-
@scottalanmiller said:
Ah, you are getting my point wrong. I never said that at all. I'm saying that you always need a copy in another place. I'm saying that you don't need to be the one that takes that copy.
I just do not believe that OneDrive (or even Amazon Cloud Drive, for that matter) would take your support call and pull something off a backup for you.
I could be entirely wring, in which case I will take a very small picture of my screen and pretend to eat it.
-
@BRRABill said:
But haven't you ever had something go wrong when it was good to have TWO backups of it?
If you are arguing that you feel that you need MORE backups than the vendor provides, that's a different thing. But if you are arguing that you need a backup, that's covered (by at least the services being mentioned here.)
If you were doing this yourself, and using a service that you created, how many backups would you have? Would you treat it the same or differently?
Example: You run your own ownCloud installation. You take a backup of ownCloud to tape via StorageCraft, which you verify daily. Or to Datto that spins it up to test, daily. Give that you have the local, the hosted and the "backup" for the same data in three places at all times (which is what you have with these services) would you then say that you need a quaternary location, which is what we are discussing here?
-
@BRRABill said:
I just do not believe that OneDrive (or even Amazon Cloud Drive, for that matter) would take your support call and pull something off a backup for you.
I could be entirely wring, in which case I will take a very small picture of my screen and pretend to eat it.
But you wouldn't need them to, right? You have your version system and you have your backups. Under what condition would you need them to pull a file off of backup for you?
-
@BRRABill said:
I am sure we've all been burned at one time or another by our friendly computers. Having a backup of the backup I think makes absolute sense.
Of course, but we have that in the use cases we are discussing, right? If I use OneDrive, as an example, I have a local copy, OD has a copy and OD takes a backup. I have three copies.
-
@BRRABill said:
How am I supposed to "version it" in OneDrive? If you mean that I should have known and renamed the files appropriately, I agree. But sometimes mistakes happens.
ownCloud offers versioning, just as an FYI. If I run ownCloud on Vultr or DO and enable snapshots for backups, I get versioning and backups of the versioning system, all in one go.
-
@scottalanmiller said:
Of course, but we have that in the use cases we are discussing, right? If I use OneDrive, as an example, I have a local copy, OD has a copy and OD takes a backup. I have three copies.
I'm talking about moving away from the local copy. Just having it in the cloud.
In a perfect world, I'd have my data local, synced to OneDrive, and also back up my hard drive to BackBlaze. Oh, and OneDrive would be cheaper.
But I am considering possibilities that leave no local copy.
-
@BRRABill said:
@scottalanmiller said:
Of course, but we have that in the use cases we are discussing, right? If I use OneDrive, as an example, I have a local copy, OD has a copy and OD takes a backup. I have three copies.
I'm talking about moving away from the local copy. Just having it in the cloud.
In a perfect world, I'd have my data local, synced to OneDrive, and also back up my hard drive to BackBlaze. Oh, and OneDrive would be cheaper.
But I am considering possibilities that leave no local copy.
Ah okay. So if you feel that you need double backups AND you don't have a local copy then yes, these services would normally be a single remote copy and a single remote backup.
-
@scottalanmiller said:
But you wouldn't need them to, right? You have your version system and you have your backups. Under what condition would you need them to pull a file off of backup for you?
Like I said, I am considering a move to the holy grail of all online.
Well, or at least trying to figure out why everyone trusts it.
-
@scottalanmiller said:
Ah okay. So if you feel that you need double backups AND you don't have a local copy then yes, these services would normally be a single remote copy and a single remote backup.
WTH have we been arguing over? LOL.
-
I'm not saying that one backup, two backups or twenty aren't correct for a given data set. Only that you should either treat the backup from your vendor as a backup or openly admit that you don't feel that you can trust them and move off of them. If you trust them but feel that more backups are needed, that's a different issue.
-
@scottalanmiller said:
I'm not saying that one backup, two backups or twenty aren't correct for a given data set. Only that you should either treat the backup from your vendor as a backup or openly admit that you don't feel that you can trust them and move off of them. If you trust them but feel that more backups are needed, that's a different issue.
I don't feel I can get to their "backups". If they mess up, sure. I think they'd be able to reproduce my data. (In what time frame, who knows.)
But I certainly don't feel like they could grab a file for me.
-
@BRRABill said:
Well, or at least trying to figure out why everyone trusts it.
Because most of us don't take many backups of our data. One main copy, one backup. That's normally what we do. And vendors like Amazon are way better at that (reliable, better equipment, etc.) than we are ourselves. Most of us don't feel that we need triplicate of soccer pictures. Nothing wrong with wanting that, but we trust Amazon or MS because, in general, they are significantly, perhaps an order of magnitude, more reliable at have a good, stable initial data store and then having a reliable, tested backup system.
-
@BRRABill said:
I don't feel I can get to their "backups". If they mess up, sure. I think they'd be able to reproduce my data. (In what time frame, who knows.)
But I certainly don't feel like they could grab a file for me.
That's my point about versioning. Tons of modern backup systems don't address file restores and rely on that to be handled "closer to the data." You fix this by introducing versioning earlier in the process.
-
@scottalanmiller said:
Because most of us don't take many backups of our data. One main copy, one backup. That's normally what we do. And vendors like Amazon are way better at that (reliable, better equipment, etc.) than we are ourselves. Most of us don't feel that we need triplicate of soccer pictures. Nothing wrong with wanting that, but we trust Amazon or MS because, in general, they are significantly, perhaps an order of magnitude, more reliable at have a good, stable initial data store and then having a reliable, tested backup system.
I'm not questioning them.
I am questioning when I save awesomesoccerphoto1,jpg, and then wife says "oh no this is the right awesomesoccerphoto1.jpg, so I overwrite it. Then says "oh not the other one was right". If I have it local, I go to BackBlaze and I'm saved. I have one copy in OneDrive ... what happens then? I tell you what. Mass anger.
I NEED TRIPLICATES OF MY SOCCER PHOTOS, LOL.
-
If you deal with backups via Digital Ocean, Rackspace, Azure, AWS, Vultr, etc. All that they offer are snapshots. Their backup mechanisms are purely around "repairing the system", not recovering an individual file. I know that some people have concerns around individual file recovery (thankfully those of us on Sharepoint don't have those worries, or those on ownCloud) and I've been lucky that in nearly thirty years of IT I've never needed to recover "a file" that I can remember. I've lost systems, but never had to go digging for a file that was lost (except for situations where people literally got rid of backups on purpose, that was a different issue and happened a LOT.)
So the culture of backups has shifted. We pretty much assume that we are going to image servers and we need to handle versions and protecting files themselves individually at a different layer. Whether you do this with GIT, SP, OC, etc. Lots of ways.
