TTL for write-ahead logs not working

Radhika Kundam


I enabled write-ahead logs to support index recovery for secondary persistence failures. I am trying to set TTL for write-ahead logs through JanusGraphManagement setting "log.tx.ttl".
Tried below use case.
1. Set write-ahead log TTL as 10 min.
2. Created few failed entries by bringing Solr(Index Client) down.
3. Waited for more than TTL time(even waited for 1hr) and bring Solr UP.
Expected behavior is failed entries should not recovered as write-ahead log might be gone by then.
Actual behavior is failed entries are recovered successfully.

I triaged and able to see that it's updating "root.log.ttl" properly while creating instance of KCVLogManager for tx log.
Please let me know if any additional configuration is required or if my understanding about expected behavior is not correct.

Thank you,