- Task
- Version · 6.0.2
- Manage
Upgrade Liquibase Secure server
Last updated: November 17, 2026
liquibase-platform update upgrades a running Liquibase Secure server deployment in place. It pulls the new images, runs the new version's database migrations, and recreates the services. Your data and secrets are kept. The server never upgrades itself and cannot schedule an upgrade, so a version changes only when you run this command.
Note: Do not use liquibase-platform install to upgrade, with or without --force. While the deployment's .env file, containers or data volumes exist, install stops with LB-SEC-INS-0011, because it cannot replace a deployment or its data.
Procedure
Troubleshooting
The upgrade stops during migration
The command stops and names the database that failed. Your data and volumes are intact, and the databases are left running so you can inspect them. Read the logs, fix the cause, then run the same liquibase-platform update command again.
./liquibase-platform logs --service dbUse --service secrets-db for the secrets database.
When the upgrade was copying the databases to PostgreSQL 17, a failure before the new version is healthy works differently. The command removes the new volumes and puts the previous version back on its old volumes, running if it was running before. With an external primary database, this ends when the migration of the primary database starts, because your server holds the only copy of that data. Fix the cause the error names, then run the same command again.
The API does not come back
The command reports that the update applied but the API is unreachable, and exits with a non-zero status. Your data is intact. Check the API and proxy logs, then the overall status.
./liquibase-platform logs --service api
./liquibase-platform logs --service nginx
./liquibase-platform statusThe upgrade stops before it changes anything
These checks run before the command stops the server. When one fails, the command stops with the code shown, and nothing in the deployment has changed:
LB-SEC-INS-0020withupdate has upgrade notices to confirm and no terminal to ask on, orLB-SEC-INS-0027: There is no terminal to ask for confirmation on. Run the command in a terminal, or add--non-interactive.LB-SEC-INS-0028: The filesystem that holds the Docker volumes has less free space than 1.2 times the size of the databases. Free space there, then run the same command again.LB-SEC-INS-0029:POSTGRES_IMAGEin.envnames an image that is not PostgreSQL 17. Set it to a PostgreSQL 17 build, then run the same command again.LB-SEC-INS-0030: Yourdocker-compose.ymlfile has local changes. The command writes the new release's file beside it asdocker-compose.yml.new. Merge your changes into that file, replacedocker-compose.ymlwith it, then run the same command again.LB-SEC-INS-0031: The external primary database runs a PostgreSQL version other than 17. Upgrade it to PostgreSQL 17 through your provider first.LB-SEC-INS-0039: The command could not read the PostgreSQL version of the external primary database. Check that the database is running and reachable from Docker, then run the same command again.LB-SEC-INS-0038: The version you passed to--tagis newer than the CLI. Run the upgrade with the CLI of that version, as in step 3.
LB-SEC-INS-0018 No installation found
update uses the directory you pass with --dir. Without --dir, it uses the directory in the LIQUIBASE_PLATFORM_DIR environment variable, then the current directory when it holds docker-compose.yml, then the directory install recorded. The error names the directory it checked and the reason, such as no docker-compose.yml. When the recorded directory no longer holds docker-compose.yml, the reason names that directory too. Pass --dir with your deployment directory, set LIQUIBASE_PLATFORM_DIR to it, or run the command from that directory. If nothing is installed yet, run liquibase-platform install instead.
Returning to an earlier version
After an upgrade to PostgreSQL 17, the old volumes still hold your data as it was before the upgrade. To go back, run the previous release's CLI with its deployment files. Its docker-compose.yml file mounts the old volumes, and anything written after the upgrade is lost. The 6.0.2 CLI does not do this for you. update with a --tag that predates PostgreSQL 17 stops with LB-SEC-INS-0036, because that release would start on the volumes from before the upgrade.
To upgrade again after you go back, the leftover PostgreSQL 17 volumes have to go first, or update stops with LB-SEC-INS-0035 and names the volume. Back it up, then remove it with docker volume rm. Remove the old volumes only once you no longer need to go back.
Otherwise, no command reverses an upgrade, so returning to an earlier version means restoring the backup you took in step 2.