On 26/01/2013 04:25 PM, Warp wrote:
> Orchid Win7 v1<voi### [at] devnull>  wrote:
>> For a home setup, I would suggest that RAID is way overkill, and a
>> simple offline backup now and then is what you actually want.
>
> The problem with this is that it's easy to forget or to just be lazy and
> procrastinate (until the day that your HD breaks.)

Yeah, well that's true enough.

(For example, I don't have any backups whatsoever. Then again, about the 
worst thing that could happen is that I lose all my photos, so...)

> The best option would be to use an external storage device and an automatic
> backupping program. This way you always have a very recent backup without
> having to do anything.

Perhaps a network storage device? IDK.

> The most typical backup program for other systems is something that you
> have to run manually, and most often than not with a horrible nightmare
> of a user interface.

I love how even extremely expensive enterprise backup systems like 
Symantec BackupExec have a horrifically awful GUI and minimal 
documentation...

Post a reply to this message

On Sun, 27 Jan 2013 09:30:06 +0000, Orchid Win7 v1 wrote:

> On 27/01/2013 02:12 AM, Jim Henderson wrote:
> 
>> rsync + cron = backup goodness.
> 
> Fine if you know your way around the CLI. Not so great for your typical
> computer user who doesn't understand all that stuff...

Fortunately, there are forums where people can ask questions about how to 
set it up. :)

Jim

Post a reply to this message

On Sun, 27 Jan 2013 05:25:46 -0500, Jim Henderson wrote:

> On Sun, 27 Jan 2013 09:30:06 +0000, Orchid Win7 v1 wrote:
> 
>> On 27/01/2013 02:12 AM, Jim Henderson wrote:
>> 
>>> rsync + cron = backup goodness.
>> 
>> Fine if you know your way around the CLI. Not so great for your typical
>> computer user who doesn't understand all that stuff...
> 
> Fortunately, there are forums where people can ask questions about how
> to set it up. :)

That is, if they can't figure out a Google search that will explain it to 
them:

http://bit.ly/WplQf2

:)

Jim

Post a reply to this message

Jim Henderson <nos### [at] nospamcom> wrote:
> On Sat, 26 Jan 2013 11:25:26 -0500, Warp wrote:

> > The most typical backup program for other systems is something that you
> > have to run manually, and most often than not with a horrible nightmare
> > of a user interface. (For example you wouldn't believe how complicated
> > it is to restore one single file with OpenSuse's system backup utility.
> > Something that takes like 5 seconds with Time Machine can take at
> > minimum 5 minutes with OpenSuse's system backup, often more. And of
> > course it's in no way automatic.)

> rsync + cron = backup goodness.

I don't think rsync does incremental backups. In other words, if you
wanted to restore an earlier version of a file, it's not possible.

-- 
                                                          - Warp

Post a reply to this message

Le 27/01/2013 11:53, Warp nous fit lire :
> Jim Henderson <nos### [at] nospamcom> wrote:
>> On Sat, 26 Jan 2013 11:25:26 -0500, Warp wrote:
> 
>>> The most typical backup program for other systems is something that you
>>> have to run manually, and most often than not with a horrible nightmare
>>> of a user interface. (For example you wouldn't believe how complicated
>>> it is to restore one single file with OpenSuse's system backup utility.
>>> Something that takes like 5 seconds with Time Machine can take at
>>> minimum 5 minutes with OpenSuse's system backup, often more. And of
>>> course it's in no way automatic.)
> 
>> rsync + cron = backup goodness.
> 
> I don't think rsync does incremental backups. In other words, if you
> wanted to restore an earlier version of a file, it's not possible.
> 
If you want to keep track of various version of a file, you need a
revision control system (and then backup the repository), such as
(<start holy war>): mercurial, cvs, rcs, svn, git, ...
( holy wars never stop, no end tag)


If hardlinks are a reality on the receiving file system, you might keep
more than one backup on the same media without doubling the actual used
size. But this starts requesting a bit of discipline.

Post a reply to this message

On Sun, 27 Jan 2013 05:53:32 -0500, Warp wrote:

> Jim Henderson <nos### [at] nospamcom> wrote:
>> On Sat, 26 Jan 2013 11:25:26 -0500, Warp wrote:
> 
>> > The most typical backup program for other systems is something that
>> > you have to run manually, and most often than not with a horrible
>> > nightmare of a user interface. (For example you wouldn't believe how
>> > complicated it is to restore one single file with OpenSuse's system
>> > backup utility.
>> > Something that takes like 5 seconds with Time Machine can take at
>> > minimum 5 minutes with OpenSuse's system backup, often more. And of
>> > course it's in no way automatic.)
> 
>> rsync + cron = backup goodness.
> 
> I don't think rsync does incremental backups. In other words, if you
> wanted to restore an earlier version of a file, it's not possible.

There are ways to accomplish that, sure.

The search I suggested to Andy, for example, has a couple of examples of 
how to do exactly that.

Jim

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:

>
> The problem with this is that it's easy to forget or to just be lazy and
> procrastinate (until the day that your HD breaks.)
>

Yeah, very true. I certainly don't back-up as often as I should. My basic
philosophy is to 'believe' in the MTBF specs of modern hard drives (probably a
mis-placed belief!) But I generate so *little* data during the week (even
usually during a month) that I don't worry too much about a loss of 'important'
data. (Except POV-Ray files!!) I do need to be more vigilant, though.

Post a reply to this message

>> I don't think rsync does incremental backups. In other words, if you
>> wanted to restore an earlier version of a file, it's not possible.
>>
> If you want to keep track of various version of a file, you need a
> revision control system (and then backup the repository), such as
> (<start holy war>): mercurial, cvs, rcs, svn, git, ...
> ( holy wars never stop, no end tag)

 From what I've seen, several revision control systems *claim* to 
support large binary files, but barf if you actually try it...

There are plenty of other ways to achieve this though. E.g., find all 
the files with the "archive bit" set, dump them into a dated folder, 
reset the archive bit on all the files you just copied. (That's why this 
file attribute exists in the first place, after all...)

Post a reply to this message

On Sun, 27 Jan 2013 17:16:16 +0000, Orchid Win7 v1 wrote:

>>> I don't think rsync does incremental backups. In other words, if you
>>> wanted to restore an earlier version of a file, it's not possible.
>>>
>> If you want to keep track of various version of a file, you need a
>> revision control system (and then backup the repository), such as
>> (<start holy war>): mercurial, cvs, rcs, svn, git, ...
>> ( holy wars never stop, no end tag)
> 
>  From what I've seen, several revision control systems *claim* to
> support large binary files, but barf if you actually try it...

subversion seems to handle them fine, FWIW.  I've seen it used to handle 
multiple revisions of multiple-gigabyte vmdk files without a problem.  
It's not terribly efficient at non-text files (it doesn't diff them, it 
just stores each revision of the entire file).

I've also seen it used for binary FrameMaker files extensively, and it 
versions those fine as well (though with the same caveat).

Jim

Post a reply to this message

Le 27/01/2013 23:28, Jim Henderson a écrit :
>>  From what I've seen, several revision control systems *claim* to
>> support large binary files, but barf if you actually try it...
> 
> subversion seems to handle them fine, FWIW.  I've seen it used to handle 
> multiple revisions of multiple-gigabyte vmdk files without a problem.  
> It's not terribly efficient at non-text files (it doesn't diff them, it 
> just stores each revision of the entire file).
> 
> I've also seen it used for binary FrameMaker files extensively, and it 
> versions those fine as well (though with the same caveat).

Yes, svn is fine... as long as no-one activate the keywords substitution
on the binary file (and a binary file can be a word document) (and hell
opens with keywords set on, when Murphy put a $Rev$ or similar in the
binary)

Post a reply to this message