Talk:Incremental backup

Sources

Is it just me or are there no sources in this article? —Preceding unsigned comment added by 219.89.133.155 (talk) 09:45, 7 June 2009 (UTC) 41.204.170.250 (talk) 15:39, 24 September 2009 (UTC)41.204.170.250 (talk) 15:39, 24 September 2009 (UTC)

41.204.170.250 (talk) 15:41, 24 September 2009 (UTC)--41.204.170.250 (talk) 15:41, 24 September 2009 (UTC)--41.204.170.250 (talk) 15:41, 24 September 2009 (UTC)Which among incremental, differential and normal backup is the most reliabe?41.204.170.250 (talk) 15:41, 24 September 2009 (UTC)--~~

Xcopy incremental

I am using this right now, and I tested it: xcopy C:\Users\Adm\Desktop\picts\*.* G:\vaio_backup\picts\ /s /h /d

The "/d" is missing from the article example, which is what makes it incremental!

Maybe because Windows is so dominant, and this command is necessary, as the 3rd part software is useless for the individual, maybe we should flesh this one out, and use it as an example of how back up scripts are made and implemented.--John Bessa (talk) 23:47, 16 August 2010 (UTC)

Opening sentence - rewrite?

I don't think the first part of the opening sentence (up to the colon) means what is intended. Apart from which its intent is not very clearly presented anyway and certainly of little use to anyone who doesn't already understand the nature of an incremental backup.

"An incremental backup preserves data by not creating multiple copies that are based on the differences in those data"

(i) *not* creating something (or multiple somethings) doesn't preserve *anything* (apart from the status quo, not backed up!); (ii) it doesn't matter what multiple copies are 'based on' if they are not created! (iii) "The differences in those data"... differences between what and what else, in what data? The only data available at that stage of the definition is the data to be preserved (actually, to be backed up... *not* 'preserved'). What relevance do the differences 'in' the, as yet, single version of the data to be backed up, have to the process of backing up. This really is a muddle of attempted expression.

There must be 100 better definitions accessible with a few minutes of Googling. Perhaps Wikipedia should use one of them.

I'm afraid this example is typical of semi-literate people writing encyclopedia entries which are then 'edited' by a random hotch-potch of 'editors', many of whom are not very literate or skilled in the use of language. Sorry if this offends, but it really is one of the metaphorical 'elephants in the Wikipedia room'. — Preceding unsigned comment added by 86.161.61.97 (talk) 19:25, 5 October 2011 (UTC)

Redirection of Differential backup

Apparently Microsoft distinguishes between these two terms. The CISSP BOK refers to how the Archive bit as managed as being the distinguishing feature of either Incremental or Differential backups. The WP article "Backup" (http://en.wiki.x.io/wiki/Backup) points out that these terms are not used consistently.

I think a proper term for the given definition, is "Incremental", since what is backed-up amounts to more storing more than the changes themselves.

FWIW. Kernel.package (talk) 23:00, 10 November 2011 (UTC)

I confused myself. "Differential" refers to a back of all changes since the last full backup, "incremental" refers to backups made since the last incremental backup (implying that the first of each shou;d incorporate the same data". Use of the Archive bit is however Microsoft defines it.

As an example, an incremental might be done daily where a differiential might be done weekly. Once the latter is done, the former backups aren't needed unless there is a need to retain a recovery point on a daily basis.

Kernel.package (talk) 23:14, 10 November 2011 (UTC)

Including Illustration

            I found the [Illustration given in "Differential Backup"] Backup to give an easy abstract of what an "Incremental Backup" is. I think it could be included here. Kaartic (talk) 14:34, 16 November 2015 (UTC)

Lots of rewriting needed

I agree with Kaartic's suggestion above, but I'd go somewhat further, maybe in a few steps. I'd expand that illustration (or table / chart) to cover two more cases, reverse incremental and reverse differential. I'd have to think a little more about what words to use in each box of the table for the reverse cases.

It would also be important to explain that, at first glance, reverse doesn't seem to make much sense, but it is used in the real world (one example off the top of my head is rdiff-backup, which uses reverse incremental). There are several advantages to this (and some disadvantages), the advantages include that the most recent backup always exists on the backup media (it doesn't have to be created (a synthetic backup) by taking an older backup and applying a differential or a series of incremental backups to it), and, because that most recent backup always exists, it is easy to make the incremental between the current content of the filesystem being backed up and the most recent backup.

Overall, the article needs work. There are quite a few things that are confusing, and some that are pretty much plain wrong. E.g., the article says: "Reverse incremental[:] An incremental backup of the changes made between two instances of a mirror is called a reverse incremental." which is misleading at best, and really, just wrong.

To be more accurate (while using too many words), you could say something like this: A backup of the changes made between two instances of a mirror "adjacent in time" is an incremental backup. (And what is meant by a mirror needs to be explained -- it is a complete backup of the system being backed up at some point in time.)

It (an incremental backup) can be either forward or reverse depending on which instance (mirror) is taken as the base (the other being the changed instance (mirror)) -- if the backup records the changes needed to make the older instance into the newer instance, it is a forward incremental, if the backup records the changes needed to make the newer instance into the older instance, it is a reverse incremental.

(Aside: I would not want to describe it in terms of two mirrors as above -- backup software doesn't compare two mirrors, but instead compares a (frozen) copy of the current content of the "live" filesystem with a mirror on the backup media. (Ok, I guess you could describe that frozen copy as a mirror, but I think that just makes the wording even more convoluted.) (But maybe image is a better word to throw into the mix.))

Maybe somewhere near the top of the article it would be helpful to define some terms, like mirror.

(Also note that it is fairly easy to convert between forward and reverse incrementals -- if something is recorded as a deletion in the forward incremental, the same thing is an addition in the reverse incremental.)

Using some of the same language as above, a backup of the changes made between two instances of a mirror "not adjacent in time" is a differential backup.

Aside: this seems to be the common terminology (distinguishing between incremental and differential), but, as noted in the Wikipedia article on differential backups, some manufacturers (not the right word) reverse that terminology.

Rhkramer (talk) 01:50, 1 August 2020 (UTC)

Cumulative incremental and differential incremental

Different source talk describe ″differential incremental″ and ″cumulative incremental″ backup. Primarily source relating to Oracle.

It would be good to include these as well.

• cumulative incremental backup is equal to differential backup
• differential incremental backup is equal to incremental backup

What do you think? Theking2 13:10, 18 April 2019 (UTC) — Preceding unsigned comment added by Theking2 (talk • contribs)

References