[jbosstools-dev] Question on data virtualization runtimes and where downloadRuntime support belongs
phantomjinx
p.g.richardson at phantomjinx.co.uk
Fri Apr 10 13:22:59 EDT 2015
On 10 April 2015 17:52:56 GMT+01:00, Rob Stryker <rstryker at redhat.com> wrote:
>On 04/10/2015 04:33 AM, phantomjinx wrote:
>>> >https://jira.jboss.org/jira/browse/JBIDE-19574 ?
>>> >
>>> >With respect for runtime detection I guess that is a completely
>separate
>>> >story.
>>> >
>>> >Is teiid a separate server like SOA-P/FSW or is Teiid something
>that is
>>> >installed into an existing EAP based server ?
>> Yes, Teiid is a module installed on top of an existing EAP server
>that provides its own API for
>> deployment of artifacts, using the jb admin port, JDBC querying of
>modelled db connections using its
>> own port (31000) and installs its own susbsytems for translators,
>data sources etc.
>>
>
>Teiid itself is something that's installed on top of an EAP server. The
>
>"Data Virtualization" product, however, comes only in an installer
>form;
>no simple unzip. So at least in terms of downloading the runtime, DV
>requires special treatment, not just an unzip. (If there's a simple
>unzip available, I haven't seen it on download manager; I've only seen
>the installer there).
>
>However you can also download "Teiid + EAP" which *does* respond to a
>simple unzip. See, for example,
>https://sourceforge.net/projects/teiid/files/teiid-8.10.0.Final-server.zip/download
>
>So while the project seems to allow a simple unzip, the product does
>not.
>
>I'm honestly OK with both code living in either place, but it's more a
>
>concern of duplication of effort / code. If I have ASTools performing
>the runtime detection, there's a few risks:
>
> 1) I may not personally stay bleeding-edge up-to-date on teiid / dv
>structural changes, what subsystems to use for detection, etc etc, and
>will require help from teiid at keeping my detection accurate
>
> 2) If ASTools is in charge of detecting, Teiid Designer (who I
>assume goes GA several months later) may find themselves limited. For
>example if a new Teiid Runtime / DV release comes out in the interval
>between our GA and Teiid's GA. Teiid Designer will either need to
>consume only what detection we provided, or, work around it by adding
>detection support for newer runtimes in their own code. Then we have
>duplication, and it becomes a split responsibility with code all over
>the place. Could get messy.
>
>I personally think it makes more sense for the detection and download
>logic to both live in Teiid Designer codebase, based primarily on these
>
>two risks. I'm aware that it limits the usefulness of ASTools
>standalone
>for users wanting to use DV / Teiid Runtime, but I also think we should
>
>be pushing people using teiid / DV to also use Teiid Designer anyway.
>
>I'll be glad to do the bulk of the work regardless where it 'lives', of
>
>course, but I just wanted to get some consensus on where the code
>SHOULD
>live first.
>
>
>
>- Rob Stryker
Little perplexed as to what code you want to "put" in Astools or Designer. As I said, the latter already detects the installed version of teiid.
As to DV v Teiid. The product of DV is EAP+Teiid wrapped up in a single installer. The layout ends up slightly different with the product having a dedicated DV directory containing the teiid modules but the result is the same.
PGR
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.jboss.org/pipermail/jbosstools-dev/attachments/20150410/99d011b0/attachment.html
More information about the jbosstools-dev
mailing list