[Info-vax] Is there a way to enabling versioning in Samba

[Stephen Hoffman]

On 2017-07-25 12:19:39 +0000, johnwallace4 at yahoo.co.uk said:

> Isn't there even an Apple-approved way of achieving the same basic 
> result? Time Machine, maybe (I don't speak Apple).

Yes.   macOS with Xcode has integrated support for DVCS including GUI 
and command-line access, and the file system provides versioning 
support without exposing the versions directly to the user as entries 
in the file system, and the default backup mechanism is Time Machine.  
Time Machine utterly blows away BACKUP for most of the common cases, 
too.   There'll be cases that Time Machine doesn't work for, but it's 
appropriate and useful for a large percentage of folks, and recovery 
from Time Machine is vastly beyond reinstallation and recovery from 
BACKUP.

As I've mentioned elsewhere, there's a lot OpenVMS can learn from other 
platforms; both good and bad.   There's a lot that people using those 
other platforms can expect of an operating system, too.

> One size does not fit all. Informal file versions seem tohave some 
> value to some customers, even if the detailshave varied over the years.

The value is largely a result of its integration in the base operating 
system — because it's there — and not because it's a particularly 
useful or well-considered or tailored implementation.   It's kinda like 
OpenVMS event flags, in that regard.   Both work fine for small and 
simple and isolated, but it eventually causes problems when scaled up 
and when mixed together in larger and more complex environments.   And 
where issues in both can tend to get ignored, until something weird 
happens in production, weeks, months or even years later.

Beyond the basic discussions of implementations, then ponder whether it 
makes sense to use a delta among the versions of many files, and also 
within the backups of those files, too.   Sure, disk space is a whole 
lot larger and cheaper than it used to be, and it'd certainly be 
possible to compress some of the data (either entirely, or as part of 
paging I/O), but why store replicated data?   And why make most folks 
even deal with PURGE and log roll-overs and the rest?  These 
capabilities are not fiction on other platforms.


-- 
Pure Personal Opinion | HoffmanLabs LLC