Description
When we commit transaction from primary to backup, we don't assign updateCounter to WAL DataRecord /Entry associated with TxEntry, altough increment updateCounter in backup partition correctly.
After that we have DataRecord's with zero update counters in WAL on backup node. This leads to WAL rebalance errors if we start suppliying process from such node partition.
We should save actual updateCounter to WAL even in case commiting to backup and introduce tests for Tx Cache WAL rebalancing.
Attachments
Issue Links
- links to