[OPEN-ILS-DOCUMENTATION] Test server for 2.1 (was: DIG Meeting Follow-up)

[Dan Scott]

On Tue, Aug 30, 2011 at 12:15:47PM -0700, Lori Bowen Ayre wrote:
> HI All,
> 
> Regarding a test server.  We've been trying to provide a test server
> environment with Brian's efforts.  Currently, I guess it is really more of a
> demo server but without too much effort, we could probably set up 2.1
> instances for libraries to use for their own testing.  Brian has the servers
> on the Amazon cloud so we could build VM instances there.
> 
> We'd probably have to charge a fee to set up a dedicated image of 2.1 that a
> library could use for a period of time.  Brian could be available to reset
> it if someone wanted to start all over as part of their testing and he could
> provide some limited tech support.  But the concept would be, your library
> could have access to their own system so you could load data and test
> settings without having to worry that another person would come in and undo
> everything you set up.
> 
> Is this of interest?  And if so, how do you see DIG being involved?  And
> also, what could libraries pay (e.g. per month) for having us set this up
> and maintain it for them....$500/month?

A VPS at Linode and many other sites with 1.5 GB of RAM - enough to load
a reasonable amount of data on for testing purposes - is about $60 /
month. Before we even go that route, though, community members have been
pretty generous in the past about making VMs available for various
purposes (PINES with the Web server and various other machines, Mohawk
with the doc server, Equinox with the testing and git servers, etc).
Maybe step one would be to ask the broader community who (if anyone)
would be willing to make a 2.1 server available.

I believe the purpose of the system would be to test and document
procedures to ensure that the documentation is sound - is that what you
had in mind, Robert and Yamil? It sounds like what Lori has in mind
would be something set up & reserved for specific libraries, which is a
bit of a different beast.

Ben Webb, as part of his Google Summer of Code project, had made
significant progress on automating Evergreen installs. So we'd be
looking at getting a 2.1 server set up, and running a single command to
reload a clean set of data whenever needed (this, too, could be
automated). It would be awesome to have a consistent set of data to
support documented task flows so we could ensure that the outcomes are
what we expect, but baby steps...

Aside: this would be a good topic for our next community meeting, which
we should probably try and schedule...