Removing telegraf data for old hosts

, , ,
1

I have not been successful in removing telegraf data collected from hosts that are no longer exist. I have reviewed numerous articles and tried everything I could find and get to execute properly, from what I can tell, and the hosts still show up in the Grafana dashboard with no data.

Steps I have taken:

  1. From the influx directory on the host, influx /#run influxdb client shell
  2. Use telegraf /# select the telegraf database
  3. SHOW TAG VALUES FROM win_system WITH KEY=host /# List all host names
  4. drop series where host=‘<for_each_stale_host_name>’ /# Remove host data
  5. show series where host=‘<for_each_stale_host_name>’ /# Confirm host name is gone
  6. Stop influx db
  7. Delete all folders under .\data\telegraf\_series
  8. influx_inspect buildtsi -datadir .\data -waldir .\wal /# Rebuild the indexes
  9. Restart influxdb

Am I missing a step? Has this worked for anyone else?

Version
Connected to http://localhost:8086 version 1.7.10
InfluxDB shell version: 1.7.8

2

You need to wait for the shards to roll over and run a compaction. Then you can do the re index

3

Thanks for the info.

Is there any way to monitor the rollover of the shards or to force them?

It looks like there is no way to initiate a manual compaction according to the docs. The only option is to do a full compaction at influxdb launch via a config file flag.

4

In enterprise there is a command truncate-shards but there is not for OSS

5

Thanks for the quick reply. Since there is not way in OSS to force shard rollover, is there a way to monitor progress to know when it is complete?