Socorro hbase cleanup

From MozillaWiki
Jump to: navigation, search

As we prepare for 2.0, a bunch of house cleaning jobs need to be run to transform the existing HBase data to be compliant with Socorro 2.0 format.

Bug: https://bugzilla.mozilla.org/show_bug.cgi?id=557827

Summary: Convert meta_data:json format from Python repr to Python JSON.
Status: Code complete, waiting for review and then deploy.

Bug: https://bugzilla.mozilla.org/show_bug.cgi?id=566340

Summary: Clean-up pre 1.7 HBase data
Status: In progress
Part 1: The new row key system uses 'salt' to spread load across region servers. Salt is basically first letter of ooid+six digits from date+ooid.
The existing rows do not contain the row-key in this format.
Convert the row-keys in 'salt' format and do value copy for all columns + families.
Some of the column values would need to be transformed/changed as detailed in Part 2.


Part 2: TBD - Daniel