The amount of metadata in the cluster makes iterating through it too slow in python. Furthermore, multiple web service processes means each process must collect its own metadata; which is also slow. Store the metadata in sqlite; queries will be faster, and data can be shared across processes.
This is still a problem
The metadata manager has been built, but it is too demanding on the cluster. In the meantime, a few explicit columns are managed, that's all.
You need to log in before you can comment on or make changes to this bug.