[CASSANDRA-16839] Truncation snapshots unnecessarily created on node startup

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 3.0.26, 3.11.12, 4.0.2, 4.1-alpha1, 4.1
• Component/s: Tool/nodetool
• Bug Category: Correctness
• Severity: Normal
• Complexity: Normal
• Discovered By: User Report
• Platform: All
• Since Version: 2.2.0 beta 1

Source Control Link

https://github.com/apache/cassandra/commit/0e12b8d4fc9183cb8bb37cb461c3fe9e434ba9b8

Test and Documentation Plan

Run CI

• Assignee: Brandon Williams
• Reporter: Paulo Motta

Authors

Brandon Williams

Reviewers

Andres de la Peña, Benjamin Lerer

• Created: 10/Aug/21 12:46
• Updated: 27/May/22 19:24
• Resolved: 11/Oct/21 17:12

Description

When testing cassandra 4.0 on ccm I noticed that everytime I restart a node, truncation snapshots are created for the tables system.table_estimates and system.size_estimates:

$ ccm create -n 1 test -s
$ ccm node1 stop
$ ccm node1 start
$ ccm node1 stop
$ ccm node1 start
$ ccm node1 nodetool listsnapshots

Snapshot Details:
Snapshot name                           Keyspace name Column family name True size Size on disk
truncated-1628599001857-table_estimates system        table_estimates    0 bytes   13 bytes
truncated-1628599099560-table_estimates system        table_estimates    0 bytes   13 bytes
truncated-1628599001736-size_estimates  system        size_estimates     0 bytes   13 bytes
truncated-1628599057438-table_estimates system        table_estimates    6.16 KiB  6.19 KiB
truncated-1628599099458-size_estimates  system        size_estimates     0 bytes   13 bytes
truncated-1628599057340-size_estimates  system        size_estimates     5.73 KiB  5.76 KiB

Total TrueDiskSpaceUsed: 0 bytes

Not sure if this is expected behavior, but feels like a bug to me.

Reproduced on 4.0, not sure if it reproduces on lower versions.