Zero-Downtime Solr Upgrade Achieved Through Direct Index Replication
Migrating search infrastructure on large Drupal sites can expose practical limits in application-layer reindexing. In a post published on 10 February 2026, Steven Jones of ComputerMinds outlines how a client running millions of indexed records moved from a legacy Solr environment to Solr 9 without downtime.
The initial plan was to stand up a parallel Solr 9 server and have Drupal’s Search API Solr module rebuild the production index. At scale, however, the reindexing process progressed too slowly to meet operational expectations.
The team instead copied documents directly between Solr cores, effectively cloning the index before switching Drupal’s configuration to the new server. By pausing write operations and maintaining read availability throughout the transition, the migration avoided visible disruption.
The case highlights a practical lesson for large installations: when index volumes are substantial, direct replication at the search layer may be faster and more predictable than rebuilding through the application stack. The full account is available on the ComputerMinds blog.


