InfluxDB-v2.6.1: How to reduce disk write load

1

Hi everyone,

I am using InfluxDB v.2.6.1 with docker. I’m having problems with high disk write loads.
I am using the database for smart-home measurements only. Therefore I’ve got a small setup with about 1.000.000 data points and six dashboards. The database size is about 120 megabytes.
Nevertheless, I am experiencing write loads of about 18 to 26 megabytes per second, which combines up to over a terabyte per day.

I already tried to reduce the write load by changing the configuration of the TSM and WAL. Sadly, this did not help at all.

Interestingly, this occurred after upgrading to v.2.6.1. I don’t know which version I ran before since I haven’t upgraded for a year or so. :smile:

Has anyone seen such behaviour before or has some idea about what is going on?
Any Help is appreciated!

2

I noticed that my InfluxDB-Instance is experiencing write errors like the following:

{"log":"ts=2023-02-18T23:06:53.014150Z lvl=debug msg=Request log_id=0g5h17Fl000 service=http method=POST host=**********:**** path=/api/v2/write query=\"bucket=weather\u0026org=grafana\u0026precision=ms\" proto=HTTP/1.1 status_code=500 response_size=90 content_length=426 referrer= remote=172.17.0.1:33090 user_agent=influxdb-client-js took=20161.127ms error=\"internal error\" error_code=\"internal error\"\n","stream":"stdout","time":"2023-02-18T23:06:53.034137955Z"}

Sadly, I can’t find a solution to this error. Has anyone an idea?