Ticket #4137 (closed enhancement: fixed)

Opened 7 years ago

Last modified 6 years ago

Deploy initial results database

Reported by: briadam Owned by: briadam
Priority: high Milestone: 5.3
Component: Software Requirements - User Version:
Severity: major Keywords:
Cc: briadam, mseldre Due Date:
Parent Tickets: #4023 Blocked By:
Blocking: Requestor:
Sensitive: no

Description

Design and implement initial version of a results database, containing information on final statistics, best points, etc. Done means this information is tracked internally and can be written to a file for user consumption.

Subtickets

Change History

comment:1 Changed 7 years ago by briadam

  • Status changed from new to accepted

From 20120103: See Acro serializer base class

comment:2 Changed 7 years ago by briadam

Developer input: In addition to being able to serialize an in-core database, need to support notion of archive-only database, where data persists only on disk. So, for example, an iterator might write it's final results, and later a print function might read them for screen output... This means the database needs to have fast on-disk lookup, insertion, and random access, not just write once, read later.

Customer input: Would like ability to directly inject into relational database, e.g., MySQL. They can currently inject the tabular data or even text output, but would want the same granularity as our new results file format.

comment:3 Changed 6 years ago by briadam

  • Status changed from accepted to closed
  • Resolution set to fixed

An initial capability for this has been deployed. Implementing it revealed numerous issues with storage of data, handling of output, and flexibility of the results data store. Will need to revisit requirements and get customer feedback, but this is done for now.

Note: See TracTickets for help on using tickets.