[CASSANDRA-11874] Schema upgrades are not guaranteed to have sequentially increasing timestamps

• Type: Bug
• Status: Open
• Priority: Normal
• Resolution: Unresolved
• Component/s: Legacy/Distributed Metadata
• Severity: Normal
• Assignee: Unassigned
• Reporter: Sebastien Bonnet
• Created: 23/May/16 10:38
• Updated: 16/Apr/19 09:30

Description

When executing schema upgrades programmatically (e.g. via the cqldriver), there is no guarantee that the upgrades will be executed in order across the nodes. This is because cassandra is always using the server generated timestamp when announcing alter table statements or other DDL changes regardless of whether a custom timestamp generator is provided when creating the session.

This could lead to failures or unexpected results. For instance

• DDL change #1 creates table A
• DDL change #2 adds a new column to table A, but finds that table A does not exist