[CASSANDRA-15833] Unresolvable false digest mismatch during upgrade due to CASSANDRA-10657

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 3.11.9, 4.0, 4.0-beta3
• Component/s: Consistency/Repair
• Bug Category: Correctness - Transient Incorrect Response
• Severity: Low
• Complexity: Low Hanging Fruit
• Discovered By: User Report
• Platform: All
• Since Version: 3.11.9

Source Control Link

https://github.com/apache/cassandra/commit/cf27558b1442e75e17e47071ecf92d1b3e5a0e36

Test and Documentation Plan

The issue can be tested by executing the following steps:

1. Start 2 node cluster of C* 3.0
2. Create a keyspace with RF=2, and a sample table with some regular columns (say of text type), insert some data into that table with CL=QUORUM or ALL
3. Flush created table on both nodes
4. Upgrade the first node either to 3.11 or 4.0
5. Start CQL, set consistency to QUORUM or ALL and enable tracing
6. Query table for a single row with a single regular column
7. Without patch, the tracing output shows the information about digest mismatch, which is unrecoverable (you can even do full repair), with the patch, there is no such issue

• Assignee: Jacek Lewandowski
• Reporter: Jacek Lewandowski

Authors

Jacek Lewandowski, Jordan West

Reviewers

Jordan West, Marcus Eriksson

• Created: 26/May/20 08:51
• Updated: 03/Jan/21 17:02
• Resolved: 28/Sep/20 15:23

Description

CASSANDRA-10657 introduced changes in how the ColumnFilter is interpreted. This results in digest mismatch when querying incomplete set of columns from a table with consistency that requires reaching instances running pre CASSANDRA-10657 from nodes that include CASSANDRA-10657 (it was introduced in Cassandra 3.4).

The fix is to bring back the previous behaviour until there are no instances running pre CASSANDRA-10657 version.

Issue Links

• causes: CASSANDRA-16148 Test failures caused by merging CASSANDRA-15833
• relates to: CASSANDRA-15946 NPE when sending REQUEST_RSP from 3.0 to 4.0 in in-jvm dtests