[DISCUSS] test results data model

[Daan Hoogland <da...@gmail.com>]

H,

I want to step away from the discussion about the cycle of releases
before we get into another deadlock there. Just let it be said that
ShapeBlue and Schuberg Philis are clearly not the same company;)

We agreed on having a central reports database of test results to be
able to generate a compatibility matrix. (citation needed) I had a
short oiliest talk with Wido about this to align it with his usage
database. When the machine for this is ready @infra we can put either
an extra schema or an extension of the usage schema to contain the
test data. I opt for the latter as the usage itself is very useful as
compatibility data as well.

a first shot at the db design:
hardware: type, version
cloudstack: version
environment: hardware.id,cloudstack.id,testdate,reportdate,contactdata

- hardware and cloudstack can be shared tables with the usage db (@widoh right?)
- I put in contact data for audibility. Usage or test data that is
anonymous shouldn't be counted as there is no way of control or
feedback.

catch22ish: contradicting reports may be submitted

thoughts?
-- 
Daan

Re: [DISCUSS] test results data model

[Wido den Hollander <wi...@widodh.nl>]

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



On 23-04-15 13:31, Daan Hoogland wrote:
> H,
> 
> I want to step away from the discussion about the cycle of
> releases before we get into another deadlock there. Just let it be
> said that ShapeBlue and Schuberg Philis are clearly not the same
> company;)
> 
> We agreed on having a central reports database of test results to
> be able to generate a compatibility matrix. (citation needed) I had
> a short oiliest talk with Wido about this to align it with his
> usage database. When the machine for this is ready @infra we can
> put either an extra schema or an extension of the usage schema to
> contain the test data. I opt for the latter as the usage itself is
> very useful as compatibility data as well.
> 
> a first shot at the db design: hardware: type, version cloudstack:
> version environment:
> hardware.id,cloudstack.id,testdate,reportdate,contactdata
> 
> - hardware and cloudstack can be shared tables with the usage db
> (@widoh right?)

Yes, correct! That should be possible. My usage reporter simply
generates a JSON which I then store in ElasticSearch.

Wido

> - I put in contact data for audibility. Usage or test data that is 
> anonymous shouldn't be counted as there is no way of control or 
> feedback.
> 
> catch22ish: contradicting reports may be submitted
> 
> thoughts?
> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
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=Hr2i
-----END PGP SIGNATURE-----