We are migrating a large production database, roughly 640gb using pgloader, from AWS RDS Mysql to AWS Aurora Postgresql. This process takes 8.5 hours, and before the instance will resolve it performs an Elasticsearch re-index that is still ongoing at +20 hours and expands to store over 100gb on disk. My team wanted to reach out too the community to ask for advice.
- Is there an easy way to shrink our database using the Sonar Admin panel options? It looks like the majority of the space is packed into just four tables: file_sources, project_measures, issues and projects.
- Is there some way to skip this initial re-index or move it into the background so we can limit our outage window?
- Alternatively, if there isn’t a way for us to skip the index, is there some way we can greatly improve performance?
Sonar v6.7.5, (We are upgrading to latest LTS after this migration)
Instance Type: c5.9xlarge: 36 vcpu, 72gb memory
Database Type: db.r5.2xlarge