We are running InfluxDB container on AWS ECS, we recently changed the index-version from inmem to tsi1, and RESTARTED the ECS container. however when I inspect from http://myinflux:8086/debug/vars, i can only see a lot node is called tsm1_* and the memory consumption does not change, looks like the configuration does not apply and influxDB still uses memory as 1st priority. Since we allocate 16GB for the container, my question is would InfluxDB only start to use disk for storing those index/cache files instead of memory, ONLY AFTER the memory limit is hit?
Related Topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
High memory usage after moving to TSI index | 1 | 2315 | January 29, 2020 | |
Tsi1 index uses more memory than without, preventing upgrade to influx2 | 2 | 521 | April 21, 2023 | |
Disk IO and memory usage on application start | 2 | 503 | June 28, 2019 | |
Influxdb 2.7.10 TSI config | 1 | 18 | September 23, 2024 | |
'tsi1' ignored? | 0 | 622 | November 27, 2017 |