[rrd-users] Architecture independence for rrd files

[Jean-Yves Avenard]

On 22 January 2013 20:30, Simon Hobson <linux at thehobsons.co.uk> wrote:
> Jean-Yves Avenard wrote:
>>I just want to do a quick resync of rrd files collecting on arm boxes
>>I have scattered around and display their content on a more beefy
>>server.
>
> Have you looked at using rrdcached ?
>
> Run cached on your central beefy box, and have your remote systems send updates to it. No need for (much) storage at the remote systems - also avoids wearing out your flash if using that instead of disk or SSD.

I did..

I quickly stopped looking further when it came to manage the
connectivity between the server and the boxes gathering the data.

The idea of setting up VPN and secured robust link on what is going to
be an unreliable connection (well, reliable enough so I could do
things like a quick rsync). The remote boxes are typically connected
via ADSL with a dynamic IP...

How does rrdcache deal with issue like the server suddenly not being
there. Is the data temporarily stored and passed on the server when it
is once again available?

I figured it would just be faster and less trouble to look at the
ability to export/import a rrd file across various architectures, but
without the need to do a dump/restore that creates huge files