Hello,
Is it possible to upgrade a two–data center Cassandra 4 deployment to Cassandra
5.0 without downtime, using a node-by-node rolling upgrade approach?
Out of curiosity, I tried adding a second data center running Cassandra 5.0 to
a test cluster based on version 4. When starting the Cassandra 5 nodes, they
fail to synchronize their schema versions.
WARN [main] 2025-11-26 06:38:09,261+0300 DefaultSchemaUpdateHandler.java:131 -
There are nodes in the cluster with a different schema version than us, from
which we did not merge schemas: our version:
(59adb24e-f3cd-3e02-97f0-5b395827453f), outstanding versions -> endpoints:
{69a38dfd-a4ba-306d-8360-4e0f91df6a0b=[cass-tmp-02-01.sandbox.lan/10.180.21.83:7000,
cass-tmp-02-03.sandbox.lan/10.180.21.85:7000,
cass-tmp-02-02.sandbox.lan/10.180.21.84:7000],
d03783d7-b468-3c1a-82f1-8e30b2edde8b=[cass-tmp-01-03.sandbox.lan/10.180.21.82:7000]}.
Use -Dcassandra.skip_schema_check=true to ignore this,
-Dcassandra.skip_schema_check_for_endpoints=<ep1[,epN]> to skip specific
endpoints, or -Dcassandra.skip_schema_check_for_versions=<ver1[,verN]> to skip
specific schema versions
