John A. De Goes
John A. De Goes
You can simulate this yourself by creating a RAM disk: http://www.cyberciti.biz/faq/howto-create-linux-ram-disk-filesystem/ We used this technique for production Precog Cloud to good effect.
There is no document, however, the labcoat-legacy repository under precog contains the version of Labcoat which was formerly used for Precog Cloud. It should work with minimal modifications. We are...
Please see #531 which tracks the status of a bundled Labcoat in a standalone Precog distribution.
This is a bug, we should find a way to avoid storing path information inside files.
Description updated with @nuttyom's feedback and #530 created for merging auth / accounts.
The repo will be updated with the new code as soon as it's ready.
What are the filters you use in queries? Those will determine the kinds of indexing we need to implement in order to accelerate this use case.
Are "date" and "jobtitle" derived attributes (computed on-the-fly in the Quirrel script) or are they present in the loaded data?
@djspiewak I think persistent indexes are required here. Either (a) the ability to dynamically add an index to some jpath(s), which is updated on ingest; or (b) the automatic indexing...