This is useful when there are datacenters with somewhat unstable connections. Previously range_repair.py stopped running if repair failed. Default behaviour is still the same (that is, fail on first), however this pull request introduces support for retries with exponential backoff.
Additional comments:
I decided to keep the retry logic in the same range_repair.py file to make it easy to copy the file to Cassandra nodes for execution.
This is useful when there are datacenters with somewhat unstable connections. Previously
range_repair.py
stopped running if repair failed. Default behaviour is still the same (that is, fail on first), however this pull request introduces support for retries with exponential backoff.Additional comments:
range_repair.py
file to make it easy to copy the file to Cassandra nodes for execution.What do you think? Tests pass (except for #37).