[Info-vax] BACKUP, rsync, Time Machine (was: Re: Re; Spiralog, RMS Journaling...)

[lawrencedo99 at gmail.com]

On Wednesday, June 22, 2016 at 2:18:18 AM UTC+12, Stephen Hoffman wrote:
> On 2016-06-20 23:55:47 +0000, Lawrence D’Oliveiro said:
> 
>> On Tuesday, June 21, 2016 at 1:13:49 AM UTC+12, Stephen Hoffman wrote:
>>
>>> On 2016-06-20 10:13:41 +0000, Lawrence D’Oliveiro said:
>>> 
>>>> rsync basically answers the question “given a source directory «src»>> 
>>>> and a destination directory «dst», what is the minimum that needs to 
>>>> be>> done to the contents of «dst» to turn it into an exact copy of 
>>>> «src»?”
>>> 
>>> Which is only part of what Time Machine does.  You're obligated to> 
>>> maintain multiple rsync targets, if you want any depth to your backups.
>> 
>> I have a script to do that. Backup run twice a day for one client, 
>> going back a week. (This is separate from the backup service provided 
>> by the central IT department.)
> 
> Good on you.   Got time-based restoration of files and directories ...

I know you’re grasping at straws, but I’ll play along...

Each backup directory is named by its timestamp, so yes.

> ... and bootable whole-disk restoration ...

That particular backup is only of user files. But yes, Linux install images normally include “rescue” functions, though the Debian one leaves off rsync. But there’s always SystemRescueCD...

> ... and fully automated hourly backup processing with increasing pruning over
> time ...

This one runs twice a day and goes back a week, but it could run hourly and go back further if you want.

> ... and pruning for target capacity?

Haven’t needed to bother with that. Just checked the backup volume, and its usage is actually slightly smaller than the one containing the original user files. As the disks gets close to full, I point that out to the client, and they put in a new set. If it’s not time to upgrade the machine anyway...

> Got the ability to store just the deltas across those 
> different output directory targets akin to a CVS ...

It does automatic file deduping, like I said.

> (akin to stacking up incrementals using OpenVMS BACKUP ...

It does not lock the data away in proprietary backup formats, thank goodness.

> ... or what most CVS packages implement internally)?

You do realize CVS is a crap VCS, don’t you? Remember the description of SVN as being “CVS being done right”? To which Linus Torvalds’ response was “anyone who says SVN is CVS done right is a moron”.

> Great!  You've managed to get rsync much closer than what 
> most folks using it have done.

It’s called “reading and understanding the documentation”. A lot of my job seems to be like that...