If layer contains more than one potential namespace, features will be
created for each namespace. Layer_feature table now has to contains one
more constrains (namespace_id).
'detector' table is added to store the metadata of detectors.
'layer_feature', 'layer_namespace', and 'ancestry_feature' tables are
modified to store the detection relationship between the
feature/namespace with the detector.
created table layer_namespace to store the many to many unique mapping of layers and namespaces
changed v1 api to provide a list of namespaces for each layer
changed namespace detector to use all registered detectors to detect namespaces
updated tests for multiple namespaces
Fixes#150
Presently the layer and namespace tables use type `varchar(128)` for
their respective name columns. For layer, this width works fine enough
using the sha256 digests provided by docker. However, if one wishes to
encode the image name into the layer name (eg, to avoid collisions like
in [0]), the limit of 128 bytes starts to feel a bit cramped. Bump to
256 bytes, since that "ought to be enough for anybody." (TM)
[0]: https://github.com/coreos/clair/issues/319
`searchNotificationAvailable` never effectively use any indexes because:
- `notified_at < $1`, where $1 is a recent timestamp, returns the
majority of the table and therefore it is cheaper for PostgreSQL
to use a sequential scan on the table.
- there is no index for `deleted_at IS NULL`.
However, when Clair has been running for long enough, the grand majority
of rows (99%+) are expected to have a non-NULL `deleted_at` field. This
commit adds a new index on this very field in order to fetch the
remaining 1% in the blink of an eye.
In other words, instead of realizing a full table scan for each
`searchNotificationAvailable` query, we'll use the small branch of a new
index, reducing the total cost from over 30k to a mere 150 on a Clair
database that has already managed more than 1 000 000 notifications.
Ales Raszka