--- title: Rebuild the TSI index description: > Use the `influxd_inspect buildtsi` command to rebuild your InfluxDB TSI index. menu: influxdb_v1: weight: 60 parent: Administration alt_links: v2: /influxdb/v2/admin/internals/tsi/rebuild-index/ --- The InfluxDB [Time Series Index (TSI)](/influxdb/v1/concepts/tsi-details/) indexes or caches measurement and tag data to ensure queries are performant. In some cases, it may be necessary to flush and rebuild the TSI Index. Use the following steps to rebuild your InfluxDB TSI index: ## 1. Stop InfluxDB Stop InfluxDB by stopping the `influxd` process. ## 2. Remove all `_series` directories Remove all `_series` directories. By default, `_series` directories are stored at `/data//_series`, however you should check for and remove `_series` files throughout the `/data` directory. ## 3. Remove all index directories Remove all index directories. By default, index directories are stored at `/data///index`. ## 4. Rebuild the TSI index Use the [`influx_inspect` command line client (CLI)](/influxdb/v1/tools/influx_inspect) to rebuild the TSI index: ```sh # Syntax influx_inspect buildtsi -datadir -waldir # Example influx_inspect buildtsi -datadir /data -waldir /wal ``` ## 5. Restart InfluxDB Restart InfluxDB by starting the `influxd` process. --- {{% note %}} ## Rebuild the TSI index in an InfluxDB Enterprise cluster To rebuild the TSI index in an InfluxDB Enterprise cluster, perform the steps above on each data node in the cluster one after the other. After restarting the `influxd` process on a data node, allow the [hinted handoff queue (HHQ)](/enterprise_influxdb/v1/concepts/clustering/#hinted-handoff) to write all missed data to the updated node before moving on to the next node. {{% /note %}}