Skip to main content

Upgrading

This guide covers upgrades for all supported deployment paths.

Upgrade Principles (All Paths)​

  • Use immutable image tags.
  • Back up database and storage before upgrading.
  • Upgrade all running Strata instances to the same image tag.
  • Migrations run automatically on startup.

1. Pre-Upgrade Backup​

Database​

PGPASSWORD=$DB_PASSWORD pg_dump \
  -h $DB_HOST \
  -U $DB_USERNAME \
  -Fc strata_production \
  > backup-$(date +%Y%m%d-%H%M%S).dump

Storage (Local Backend)​

docker compose cp web:/rails/storage ./storage-backup-$(date +%Y%m%d)

If you use object storage (STORAGE_BACKEND=amazon|google|microsoft), use your provider backup/versioning controls.

2. Upgrade by Deployment Path​

Quick Install / Single-Host Docker Compose​

cd /path/to/strata
curl -fsSL https://strata.do/server/install.sh | bash

The installer preserves .env, updates compose config, pulls the selected image tag, and restarts services. Re-running the installer renames the background service from worker to job in docker-compose.yml (stop the old stack first if you have a running worker container).

Manual Host Deployment (VM / On-Prem)​

  1. Set target image tag in .env:
STRATA_VERSION=1.0.0
  1. Pull and restart:
docker compose pull
docker compose up -d
  1. Verify:
docker compose logs -f web job
curl http://<your-host-or-alb>/up

AWS ECS Manual Deployment​

  1. Select target Strata image tag.
  2. Create new task definition revisions for web and job services.
  3. Update the web ECS service first, then the job service (force new deployment).
  4. Verify ALB /up and application logs in CloudWatch for both services.

3. Rollback​

Docker Compose (Quick Install / Manual Host)​

  1. Set previous STRATA_VERSION in .env.
  2. Restart:
docker compose pull
docker compose up -d
  1. If required, restore database and storage backups.

AWS ECS​

  1. Revert the ECS service to the previous task definition revision.
  2. Verify ALB /up and application logs.
  3. If required, restore database and storage backups.

4. Multi-Node and HA Notes​

  • Upgrade in rolling order, keeping at least one healthy instance when possible.
  • Keep STRATA_SECRET_KEY_BASE and all Strata encryption keys identical across all nodes.
  • Ensure only one migration writer at a time (via normal rollout sequencing).