[SAC] Ad hoc VM Proposal

Folks,

We had an extensive discussion on IRC today about some conflicting needs.

The Mapbender projects would like somewhere they can host development
versions of Mapbender to show new capabilities and to facilitate development.

It was discussed that we might put this on the ProjectsVM, but it was felt
that this might put the whole VM at risk as "not ready for prime time"
code would be running on a server with other relatively mission critical
services (like project web sites).

It was also discussed that we might provide one or more telascience blades
for such purposes for Mapbender or other projects. This is plausible, though
currently we are a bit stymied with getting them reinitialized.

We also discussed creating a project specific VM for the Mapbender project,
as was done for QGIS. But it was felt that it set a bad precident to
provide project specific VMs for what seems like a less than compelling
need particularly since we can't afford to issue most projects their own
VMs on our available hardware. The resources are not sufficient.

The compromise Howard, Chris and I came to was to establish an "ah-hoc"
VM at OSU OSL which anyone in the telascience ldap group (ie. anyone who
can login to the Projects VM or the blades) could use for development,
and demonstration purposes with the understanding that the VM might not
be all that stable. SAC would not provide any special support on this
VM, and sudo access would be provided fairly freely (while we are trying
to be cautious about that on the Projects VM).

Does anyone in SAC have an objection to such a VM being setup? If I don't
hear any stringent objections, I'll likely try and issue a request to OSU OSL
soon. Once created, I'll take care of initial setup and documentation in the
wiki.

Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, warmerdam@pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Programmer for Rent

On Jul 26, 2010, at 1:26 PM, ext Frank Warmerdam wrote:

Folks,

<snip>

Does anyone in SAC have an objection to such a VM being setup? If I don't
hear any stringent objections, I'll likely try and issue a request to OSU OSL
soon. Once created, I'll take care of initial setup and documentation in the
wiki.

I feel that this is a solid balance between the needs of developers and our
current resources.

Thanks,
--
Christopher Schmidt
Nokia

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Frank Warmerdam wrote:

Folks,

We had an extensive discussion on IRC today about some conflicting needs.

The Mapbender projects would like somewhere they can host development
versions of Mapbender to show new capabilities and to facilitate
development.

It was discussed that we might put this on the ProjectsVM, but it was felt
that this might put the whole VM at risk as "not ready for prime time"
code would be running on a server with other relatively mission critical
services (like project web sites).

It was also discussed that we might provide one or more telascience blades
for such purposes for Mapbender or other projects. This is plausible,
though
currently we are a bit stymied with getting them reinitialized.

We also discussed creating a project specific VM for the Mapbender project,
as was done for QGIS. But it was felt that it set a bad precident to
provide project specific VMs for what seems like a less than compelling
need particularly since we can't afford to issue most projects their own
VMs on our available hardware. The resources are not sufficient.

The compromise Howard, Chris and I came to was to establish an "ah-hoc"
VM at OSU OSL which anyone in the telascience ldap group (ie. anyone who
can login to the Projects VM or the blades) could use for development,
and demonstration purposes with the understanding that the VM might not
be all that stable. SAC would not provide any special support on this
VM, and sudo access would be provided fairly freely (while we are trying
to be cautious about that on the Projects VM).

Does anyone in SAC have an objection to such a VM being setup? If I don't
hear any stringent objections, I'll likely try and issue a request to
OSU OSL
soon. Once created, I'll take care of initial setup and documentation
in the
wiki.

Best regards,

Frank,
thanks for this excellent summary and we are more than happy to go down
this road and see how it works out.

Please let this list know when the machine will be ready and how to
access it. Ideally it would also authenticate against the OSGeo LDAP so
that we can use the same credentials.

Thanks a lot,
Arnulf

- --
Arnulf Christl

Exploring Space, Time and Mind
http://arnulf.us

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkxN/v0ACgkQXmFKW+BJ1b1uUwCfV84ugBKr2jf//8BW7XUeXbg7
RJ8AnjjjrQ+y4cE/Ivgwz/0drcLOWTNq
=qk6H
-----END PGP SIGNATURE-----

Frank:

> VM at OSU OSL which anyone in the telascience ldap group (ie. anyone
> who can login to the Projects VM or the blades) could use for
> development,

Arnulf:

Ideally it would also authenticate against the
OSGeo LDAP so that we can use the same credentials.

(telascience is an access group within the OSGeo LDAP setup, no new
credentials req'd)