Re: Digest mismatch

Thank you.  OK - I can see from 'nodetool getendpoints keyspace table key' that 3 nodes respond as one would expect.  My theory is that once I encounter the error, a read repair is triggered, and by the time I execute nodetool, 3 nodes respond. I tried a test with the same table, but with LOCA

Antwort: Re: Increased read latency with Cassandra >= 3.11.7

Thanks Paulo! I'll give it a try and let you know. -"Paulo Motta" schrieb: - An: user@cassandra.apache.org Von: "Paulo Motta" Datum: 03.12.2020 13:39 Betreff: Re: Increased read latency with Cassandra >= 3.11.7 As a workaround if your TWCS table is append-only (no updates or deletions

Re: Increased read latency with Cassandra >= 3.11.7

As a workaround if your TWCS table is append-only (no updates or deletions) you could try enabling unsafe_aggressive_sstable_expiration (see https://cassandra.apache.org/doc/latest/operating/compaction/twcs.html) so inactive sstables will be expired faster and will not be included during reads. Em

Re: Increased read latency with Cassandra >= 3.11.7

I think this could've been caused by https://issues.apache.org/jira/browse/CASSANDRA-15690 which was introduced on 3.11.7 and removed an optimization that may cause a correctness issue when there are partition deletions. I'd suggest you to open an issue at https://issues.apache.org/jira/projects/CA