Stale volume data causes upgrade failures

critical

storageUpdated Feb 16, 2026(via Exa)

Sources

Troubleshooting | getsentry/self-hosted | DeepWikideepwiki.com

Technologies:

ClickHouseThe root cause of this issue originates in ClickHouse

PostgreSQLThe root cause of this issue originates in PostgreSQL

Apache KafkaThe root cause of this issue originates in Apache Kafka

RedisThe root cause of this issue originates in Redis

How to detect:

Upgrades fail when Docker volumes contain incompatible data from previous versions. Affects critical volumes: sentry-postgres, sentry-clickhouse, sentry-kafka, sentry-redis, sentry-zookeeper (legacy). ClickHouse upgrade must run first while old containers are still running.

Recommended action:

Stop all services with 'docker compose stop -t 60'. Remove problematic volumes with 'docker volume rm sentry-self-hosted_<volume-name>'. Re-run installation with './install.sh'. Critical: ensure ClickHouse upgrade runs while old containers are running before stopping services.