bdbeeb992d
two major pieces: * use context_type/context_id instead of context_code for three relations so that sharding extensions will automatically work * deserialize stream item data as actual AR objects instead of OpenObject (for similar reasons) test plan: * run the predeploy migration, but not the postdeploy one * view your dashboard on a shard other than your home shard; it should be identical to your home shard * view courses that you are enrolled in with the same id but on different shards. the activity stream should have the correct data for the course in that shard, for both courses (i.e. no mixing of data from the two shards) * do actions that generate new stream items, and verify the new items appear correctly for the above two steps * run the postdeploy migration, wait for the job to finish, then repeat the test plan Change-Id: I8d5fcadb8d971acf7388a12e9151a3e927751f44 Reviewed-on: https://gerrit.instructure.com/15462 Reviewed-by: Jacob Fugal <jacob@instructure.com> Tested-by: Jenkins <jenkins@instructure.com> |
||
|---|---|---|
| .. | ||
| stream_item_cache_spec.rb | ||