Upgrade/Installation takes long for large databases

[patrickatamaniuk]

platform: 2.0.4, 2.1.6, 2.2.1 # supported: 2.2.4, 2.3.2
backlogs: 0.9.35 # supported: 1.0.5
ruby: 1.9.3, 1.8.7 # supported: 1.9.3, 2.0.0

Upgrades from 0.9.26 and prior as well as installing backlogs on existing databases can take quite long.
DB with more than 50.000 issues takes more than 3 days.

Workaround: If historic burndown graphs are not relevant, perform the installation/upgrade with the environment variable
rbl_migration_ignore_historic_history=yes

[patrickatamaniuk]

This should be noted in the installation instructions?

Ideas:
Maybe the rebuild process can be optimized? only upgrade stories/tasks if bl is configured? Restrict upgrade to bl-relevant tracker types?
Evaluate, if the rebuild rake task can run concurrently to an online production system (using ionice and stuff)

[patrickatamaniuk]

possible solution at #806