Are All Copies Backups?



  • This question came up in another thread and I figured that it was worth a discussion:

    If you have an unintentional copy of data, does that constitute a backup? Could you recover from said copy? Yes, of course. But are all copies worthy of being considered a backup? In some ways, I can see how the terms could be basically interchangeable (when a copy is fully decoupled, of course.)

    But, for example, do you consider the copy of data sitting in a cache to be a backup? I don't, it is intentionally transient. But there is a brief moment in time when it could be used for a recovery. But just because you can recover does not really mean that you had a backup, only that you found a means of recovery.

    I feel like the concept of a backup requires a certain degree of intention. A copy is a copy, no matter why it exists or how you are attempting to use it. But a backup, I feel, implies an intentional to retain for the purpose of protection against loss.

    Thoughts?



  • I've always classed backup as a conscious effort to make a copy of something for the sole purpose of being able to recover it if it all goes south...accidentally duplicating/storing in cache etc in my opinion isn't a backup as it isn't reliable...but then, some "backup solutions" can be far from reliable too!!



  • @NattNatt said:

    I've always classed backup as a conscious effort to make a copy of something for the sole purpose of being able to recover it if it all goes south...accidentally duplicating/storing in cache etc in my opinion isn't a backup as it isn't reliable...but then, some "backup solutions" can be far from reliable too!!

    I'd not go so far as to quantify reliability in the definition. Intent is black or white. Either you intend to retain or you don't intend to retain. You could add a third possibility... you intend to retain, you don't intend to retain or you intend not to retain.

    I feel that only the first option qualifies as a backup. The other one (or two) would just be copies.



  • I wouldn't consider a copy a backup unless you know it's there, and know how you can recover from it.

    For example, the iTunes iPhone Backup software.

    This is clearly a backup tool designed by Apple for their product line. But many of the user base has a hard time remembering about it. So for all intensive purposes those users have no idea that "I have a backup". Unless if they have ever connected their phones to iTunes with the intent of making said backup of their device.



  • The question came up in the situation where my mobile phone retains a copy of a file that I have already transferred its primary storage to another location but has not yet had the deletion of that file completed. Is it "backed up" to the phone? I say no, the file may exist there but the intention is to delete it. In my case, it's not even the "not intending to retain" case but the more specific "intending not to retain" case. To me, that's no backup.



  • @DustinB3403 said:

    I wouldn't consider a copy a backup unless you know it's there, and know how you can recovery from it.

    I would not go this far because it creates some serious complications. Like backups becoming ephemeral based on the knowledge of them and the knowledge of how to recover them. Meaning something being backed up depends on the state of the observer. Your backup is my copy, for example.

    The intention concept is not ephemeral, was there an intent to protect the data at time of the action by the person initiating the action. But needing to know that a backup exists and how to use it would make the backup like Schroedinger's Cat - you can't know if there is a backup until it is no longer a backup and the only answer could ever be that there "was" a backup.



  • I tend to agree with the intent thinking here Scott - and even though I was using the term backup in that other thread, I currently lack a more accurate term. Additionally, the primary source of a data piece vs the backup or a synced copy would depend upon which is the one in use at the moment.

    Phones bring two things to easy conversation - photos and contacts.

    With photos we primarily consider them static, once they are taken they don't change, they just get copied around from device to device.

    So we take a picture on our phone, it saves to the local storage. When the phone has a network connection it copies that picture to the cloud storage. At which point does the local picture change from being the primary to being the backup?

    Now contacts. Contacts primary source is dependent upon where they are being edited at the time. A new blank phone would be the backup/synced copy of Exchange's contact list. But once you edit a contact on the phone, which is done entirely locally first - that is now the primary, at least until it's sync'ed back to Exchange.

    It is a lot of semantics.



  • @scottalanmiller said:

    The intention concept is not ephemeral, was there an intent to protect the data at time of the action by the person initiating the action. But needing to know that a backup exists and how to use it would make the backup like Schroedinger's Cat - you can't know if there is a backup until it is no longer a backup and the only answer could ever be that there "was" a backup.

    I wouldn't look at it like this, lets use the iPhone example.

    When you connect your iPhone to iTunes it clearly asks if you would like to make a backup of "this iPhone" to the Apple Cloud or to the local computer.

    If you intentionally connected the phone to the computer to run this process, that is clearing building a backup. It exist because you "made it". The cat is clearly in the box. Now to test it (as you would with any backup) you'd attempt the restore to your iPhone. Which now means the cat is alive in the box.



  • Which is a backup even a backup if it's not confirmed that the cat is alive? IE If you can't restore from your "backup" does it qualify as being a backup at all or just a failed attempt at a backup?



  • @Dashrender said:

    I tend to agree with the intent thinking here Scott - and even though I was using the term backup in that other thread, I currently lack a more accurate term. Additionally, the primary source of a data piece vs the backup or a synced copy would depend upon which is the one in use at the moment.

    Phones bring two things to easy conversation - photos and contacts.

    With photos we primarily consider them static, once they are taken they don't change, they just get copied around from device to device.

    So we take a picture on our phone, it saves to the local storage. When the phone has a network connection it copies that picture to the cloud storage. At which point does the local picture change from being the primary to being the backup?

    Now contacts. Contacts primary source is dependent upon where they are being edited at the time. A new blank phone would be the backup/synced copy of Exchange's contact list. But once you edit a contact on the phone, which is done entirely locally first - that is now the primary, at least until it's sync'ed back to Exchange.

    It is a lot of semantics.

    Just because it is edited locally does not make that the primary, IMHO. For example, do you feel that MangoLassi's primary database is the one that GroveSocial hosts or the one running inside your browser, making ML a multi-master system.

    In case you are wondering, ML creates a NoSQL database in memory in your browser and your changes all go there first before syncing up to the main MongoDB system. Like contacts on your phone before going to Exchange.



  • @DustinB3403 said:

    Which is a backup even a backup if it's not confirmed that the cat is alive? IE If you can't restore from your "backup" does it qualify as being a backup at all or just a failed attempt at a backup?

    All systems fail. When the government demands that you maintain a backup, they do not require that it be successful. You need something that you can say "yes we have a backup." If the term isn't backup, then would need a new one and when could you use the term "backup"?

    For example, if you require absolute success for backup to be the term, then you can never tell someone to take a backup or say that you have a backup or that you retain backups. While I understand your point, is a backup that might not work really a backup, it would make the term effectively unusable.



  • @scottalanmiller Well Scott would you take more then a second thought at a failed backup as being usable, or simply say "you have no backup because it's unusable."

    The backup process failed, if it's unusable, so you have to create another backup, and test the backup.

    Once you have a successful backup, that can be used to restore from, then you have a backup.



  • It's the intent of "I have something to revert to or recover with should something go south." Not "Oh I made this so I'm in good shape without testing it"

    That is just being blind about the entire situation and the goal, you need to confirm that the backup "works" otherwise what's the point if you have no way to recover?



  • @DustinB3403 said:

    It's the intent of "I have something to revert to or recover with should something go south." Not "Oh I made this so I'm in good shape without testing it"

    That is just being blind about the entire situation and the goal, you need to confirm that the backup "works" otherwise what's the point if you have no way to recover?

    I definitely see where you are going with this, but there are situations where testing every backup is just not feasible. You do random tests on backups, but you don't test every single one - are you saying the non tested backups aren't backups? What about tested backups that are on media where the media fails after the test? Is that still a backup ? I think so.



  • @Dashrender My point is, you're building a backup for a purpose, and to ensure that the backup works, you test it.

    That is what makes it a backup.

    Sporadic testing, is still testing. But if you have a single "backup" that is broken and just doesn't work it's not a backup, because you can't recover from it.



  • @DustinB3403 said:

    That is what makes it a backup.

    nooo... is it smart to test it? yes, but I don't agree that testing is intrinsic to it being a backup.

    It's like Scott's earlier post

    @scottalanmiller said:

    I'd not go so far as to quantify reliability in the definition. Intent is black or white. Either you intend to retain or you don't intend to retain. You could add a third possibility... you intend to retain, you don't intend to retain or you intend not to retain.

    Your testing requirement is the same as the reliability quantification that NattNatt was looking to add.



  • Now I will add - if your backup process fails - then you clearly Don't have a backup. But if the process claims that it is successful, well then I say generically you do have a backup, wither it's reliable/usable isn't known until you try using it.



  • @Dashrender said:

    So we take a picture on our phone, it saves to the local storage. When the phone has a network connection it copies that picture to the cloud storage. At which point does the local picture change from being the primary to being the backup?

    I say when it hits the cloud. Because the cloud is your storage. And the cloud provider has a backup. (As @scottalanmiller often says WE are the IT user in that case, not the IT Admin. That is the job of the cloud provider.) The copy on your phone is just synced, must the same as your contacts.

    Me? I like a little more granular control over my data than just hoping the cloud provider can get it back. Even Apple recommends downloading your library somewhere other than iCloud.



  • @Dashrender But why would the process reportedly succeed, but still not work. If it doesn't work, you have nothing but wasted resources.

    You have no backup at that point.



  • @BRRABill said:

    Me? I like a little more granular control over my data than just hoping the cloud provider can get it back. Even Apple recommends downloading your library somewhere other than iCloud.

    LOL even they don't trust their ability to restore your data. iCloud isn't O365. Hell OneDrive isn't O365. I wouldn't expect either service to have any onus toward recoverability of my data unless it's specifically in their TOS.



  • @Dashrender said:

    LOL even they don't trust their ability to restore your data. iCloud isn't O365. Hell OneDrive isn't O365. I wouldn't expect either service to have any onus toward recoverability of my data unless it's specifically in their TOS.

    Considering my recent (and andgoing) issue with iCloud, I am starting to get a little nervous. 🙂



  • TL:DR rest of thread:

    A copy is a copy,.. I can have 30 copies of a document,... but the one I'm working on could get lost or damaged. But I haven't copied back to the other 29 locations,.. so I've lost hours of work.

    A back up - a true back up looks at the original.. Better to have 20 backups theAn 20 copies.

    And a good back up is only as valid as the last tested restore.

    Yes @NattNatt ----thAn; speed typing... smh



  • @gjacobse "Better to have 20 backups Than 20 copies" I assume you mean?


  • Vendor

    I think of copies and backups like this: a backup is a conscious effort to protect your data, while a copy is just a duplicate of your data. Backups are most likely run on a regular schedule, and offer some way to easily restore the data when needed, as well as check to make sure that they're working correctly..You may be diligent enough to copy your files after each change, but most likely,you're not. If you don't make a new copy with each change, or even once daily, how useful is a copy if you need to restore from it? Also, how do you check to ensure that the copy hasn't been corrupted?



  • So this is computer existentialism?



  • I think you're getting two separate things mixed up needlessly.

    1. Intentional backups / backup strategy

    2. Potential sources of data to recover from

    Bottom line, if you're in a position to need to recover stuff it makes lots of sense to look at both 1 and 2. Should you depend on 2? Hell no, that's the whole point of 1.


  • Banned

    To me.

    A Backup is a read only copy that can not be alterted.. Otherwise it's living data, And you can not verify the data has not changed/added or removed since the backup was taken.


Log in to reply