[GLLUG] portability and power

[James Courtier-Dutton]

On 22 February 2014 07:50, JLMS <jjllmmss at googlemail.com> wrote:
>
> On 22 Feb 2014 07:01, "John G Walker" <johngeoffreywalker at yahoo.co.uk>
> wrote:
>>
>>
>>
>> On Fri, 21 Feb 2014 23:37:17 +0000 Justin Perreault
>> <justinperreault at dl-jp.com> wrote:
>>
>> > I would like a set up where I have 2-3 computers running the same
>> > software and containing the same data files. I have been considering
>> > rsync, csync, virtualisation, version control systems, servers...
>> >
>> > All systems are potentially mobile and could be without network
>> > connectivity. There is also potential for only one system to be
>> > available to all users. They need to operate individually as well as
>> > syncing files from each when connected.
>>
>> I use Dropbox for this. Is there any reason why you can't use this
>> solution, or create a setup that runs n the same principle (ie holds
>> data in the cloud and updates every time it can connect)?
>>
>> --
>>  All the best,
>>  John
>>
>
> It is called a file server :-)
>
> Dropbox, cloud ... NFS, Samba... NAS... One doesn't go replicating data all
> around the place, one uses a file server.
>
> As for cloning systems ( a logical separation should be made between your
> data and your software) create a master copy of your system and document a
> procedure to do the cloning you are comfortable with.
>
> Machines going off line should suck all the data they need from the file
> server ( assuming this happens in a controlled manner).
>
> As for synchronizing data when a machine goes back online it depends
> entirely in how your data is organized, one would need to know more in order
> to suggest something.
>
>

A "file server" might be the solution.
The problem seems to me to be more of a MDM (Master Data Management)
type solution.
You need to decide where the master version of each document or piece
of data is stored.
Everyone else should refer to the master or make copies of the master.
Once you have that decided, you can then work out a solution to how
one goes about controlling/restricting the modifying of that master
data.
E.g. If two people try of edit the same file/data at the same time,
what should happen?

A "file server" if often a good solution for file management, because
you know that the "master" version is stored on the file server, and
the file server does file locking, so only one user can modify a file
at the same time.

A "file server" is not always the correct solution.
You might wish to keep a record of all previous versions of a
particular file, in which case a repository might be needed. E.g. git
or svn.
Alternatively, you could use btrfs filesystem level version tracking
and do snapshots. Then use btrfs send to sync the file-system between
computers.
You might wish two or more people to be able to edit the same
file/data at the same time, in which case some collaboration
middle-ware might be needed.
There are many different solutions. You just need to think about the
problem at first from a "requirements" perspective, and not an already
designed solution perspective.

Kind Regards

James