Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Later
-
None
-
None
-
None
Description
I have been involved in a case where user of verifyrep observed fluctuating row key count during successive runs with time range.
If verifyrep can persist the row keys during current run (to e.g. hdfs) (controlled by a flag), successive run(s) would be able to use this information to find the difference in row keys w.r.t. previous run and perform some actions (such as performing raw scan, etc).
This would allow the user to correlate results from successive runs better.