High load average, low CPU usage when starting InfluxDB 2.x via docker. No query / write to DB

When i start influxdb via docker, immediately the load-average starts to build up (enormously).
This happened as a consequence of me writing a lot of data to the DB (I guess).

Here is what I did:

  1. I started the DB like so: docker run --restart unless-stopped -d -v /path/to/dir:/path/to/dir --name influxdb -p 8086:8086 quay.io/influxdb/influxdb:v2.0.1 --reporting-disabled --bolt-path=/path/to/dir/.influxdbv2/influxd.bolt --engine-path=/path/to/dir/.influxdbv2/engine
  2. I started making about 500 writes per second. The average 15 min load on a 8 core CPU was about 4. Ram was at about 15% and CPU usage was at about 15%.
  3. An hour into this the load started to sky rocket to values in the 60s - 100s.
  4. I stopped all processes (and disabled) and rebooted the machine.
  5. On start I had nothing running and only started the docker instance of the DB.
  6. The load started increasing again despite me not writing and data to the DB. (see picture attached below).

Can someone please help me debug this?

The error was caused because I was heavily writing data on the side (and forgot about this).
What was interesting is that the python-client persistent in writing data even when re-connecting the DB - i.e. the client retired on failures, which I did not expect as behaviour. I was expecting the python client to fail if the DB is restarted.

1 Like