2 months ago
I upgraded MySQL from 9.4 to 9.6 and the upgrade failed on internal table mysql.slave_master_info.
Now MySQL cannot start because the server upgrade fails.
The volume is still mounted and data is intact.
I need help mounting the volume in a safe mode or extracting a dump from the existing volume without running the upgrade process.
6 Replies
2 months ago
Your MySQL service in the acm-suite-migration project is currently stuck in a DEPLOYING state. Since the volume data is intact, your best recovery path is to restore a volume backup from before the upgrade attempt - check the Backups tab on the MySQL service to see if any backups are available, and if so, restore one from before the 9.6 upgrade. After restoring, roll back the deployment to the previous MySQL 9.4 image via the Deployments tab (click the three dots on the last successful 9.4 deployment and select Rollback). If no backups exist, you can deploy a separate temporary service using the MySQL 9.4 image, mount the same volume to it (or a copy), and use it to extract a dump before attempting the upgrade again.
Status changed to Awaiting User Response Railway • about 2 months ago
Status changed to Awaiting Railway Response Railway • about 2 months ago
Railway
Your MySQL service in the acm-suite-migration project is currently stuck in a DEPLOYING state. Since the volume data is intact, your best recovery path is to restore a volume backup from before the upgrade attempt - check the Backups tab on the MySQL service to see if any backups are available, and if so, restore one from before the 9.6 upgrade. After restoring, roll back the deployment to the previous MySQL 9.4 image via the Deployments tab (click the three dots on the last successful 9.4 deployment and select Rollback). If no backups exist, you can deploy a separate temporary service using the MySQL 9.4 image, mount the same volume to it (or a copy), and use it to extract a dump before attempting the upgrade again.
2 months ago
Please attach my existing volume to a MySQL 9.4 instance without triggering upgrade.
2 months ago
Hello
2 months ago
I need help recovering data from the mysql-volume volume of the MySQL-PB4K service.
I've confirmed the following:
The original mysql-volume was moved to the MySQL-PB4K service and is mounted at /var/lib/mysql.
With MySQL 9.4, the service fails due to an invalid downgrade: "Cannot downgrade from 90600 to 90400".
With MySQL 9.6, the service starts InnoDB but fails during the server update on the internal table mysql.slave_master_info with "Failed to delete SDI" and then aborts.
I need you to mount this volume on a rescue instance or container to extract a logical dump of the user databases, or provide me with the procedure supported by Railway to export the data without running the broken update.
Railway
Your MySQL service in the acm-suite-migration project is currently stuck in a DEPLOYING state. Since the volume data is intact, your best recovery path is to restore a volume backup from before the upgrade attempt - check the Backups tab on the MySQL service to see if any backups are available, and if so, restore one from before the 9.6 upgrade. After restoring, roll back the deployment to the previous MySQL 9.4 image via the Deployments tab (click the three dots on the last successful 9.4 deployment and select Rollback). If no backups exist, you can deploy a separate temporary service using the MySQL 9.4 image, mount the same volume to it (or a copy), and use it to extract a dump before attempting the upgrade again.
2 months ago
I need help recovering data from the mysql-volume volume of the MySQL-PB4K service.
I've confirmed the following:
The original mysql-volume was moved to the MySQL-PB4K service and is mounted at /var/lib/mysql.
With MySQL 9.4, the service fails due to an invalid downgrade: "Cannot downgrade from 90600 to 90400".
With MySQL 9.6, the service starts InnoDB but fails during the server update on the internal table mysql.slave_master_info with "Failed to delete SDI" and then aborts.
I need you to mount this volume on a rescue instance or container to extract a logical dump of the user databases, or provide me with the procedure supported by Railway to export the data without running the broken update.
Status changed to Open brody • about 2 months ago
2 months ago
The recovery container can mount the volume, and InnoDB initializes, but MySQL still fails with Data Dictionary initialization failed. When I use --upgrade=NONE, the server explicitly reports that upgrade is required but prohibited by that option. Without that option, it attempts the 9.4 → 9.6 upgrade and aborts during system table migration.
This suggests the volume is stuck between a partially upgraded state and a required data dictionary upgrade path. At this point, I need Railway assistance to either:
- restore a backup/snapshot from before the upgrade, or
- mount the volume in an internal recovery environment and extract a logical dump of user databases.