Compactor not working after import of historical data from influxdb2

Dear All!

I am using influxdb2 for few years together with openhab for my home automation.

I upgraded my Synology NAS and in this step I want to upgrade also the influx database to V3.

Influxdb3 is running on my NAS (with 32GB memory) as a Docker image.

But then I ran into several problems importing my old data from influxdb2 …

What I did:

  • First I tried to exported it into a big lp.gz file (about 240MByte (from 2022 …)); this was one of my first try, but it is not working on my NAS with limited (home license) cpu/memory.

    So I exported the data via script day by day to a lp file (from the old influxdb2) …

    influxd inspect export-lp \
    –bucket-id “$BUCKET_ID” \
    –engine-path “$ENGINE_PATH” \
    –start “$start_ts” \
    –end “$end_ts” \
    –output-path “${OUTPUT_DIR}/export_${current_date}.lp.gz” \
    –compress

    Working well, no problems.

  • I tried importing it several times, but I encountered multiple issues (memory, hanging …) so I made my last attempt:

    • I wrote a smal script to merge the daily files into monthly files

    • I setup the memory threshold to about 250MByte - INFLUXDB3_FORCE_SNAPSHOT_MEM_THRESHOLD=256
      For less imported data, the compressor should be able to work more easily …

    • The I imported it in chunks of about 80000 lines (INFLUXDB3_MAX_REQUEST_SIZE_BYTES=104857600) …, to get larger L1 files !
      export_2022-71.lp to export_2023-05.lp (about 216MByte of lp lines), then I stoped the import to check the compaction process!

    • This wrote me L1/GEN1 parquet files (enterprise version) without problems …, about 227000 parquet files!

    • But if I look at the compactor I found many plans but only few run_completed …

    • In the meantime I switched from the home license temporarly to the trial licens to have more cpu’s available, but it doesn’t help.

  • I have tried several times to start over with changed settings, but now I have no idea what the actual problem is.
    It looks like the compactor is not selecting the old data sets for compression.

Setups/logs:

  • influxdb3 show nodes --token apiv3_
    | node01 | 0 | 30b7abd4-545d-4926-b82a-d4264b6d0991 | [all] | 6 | running | 2026-05-01T17:50:24.255 | 063b9e0efa36:8181 | {“exec-mem-pool-bytes”:“8000000000”,“num-datafusion-threads”:“4”,“license-email”:“*******”,“gen1-lookback-duration”:“5y”,“mode”:“all”,“license-type”:“trial”,“query-file-limit”:“1000”,“parquet-mem-cache-size”:“100”,“retention-check-interval”:“24h”,“table-index-cache-max-entries”:“200”,“serve-invocation-method”:“docker-hub”,“cluster-id”:“cluster01”,“object-store”:“file”,“last-cache-eviction-interval”:“1m”,“datafusion-max-parquet-fanout”:“50000”,“compaction-gen2-duration”:“20m”,“plugin-dir”:“/var/lib/influxdb3/plugins”,“compaction-max-num-files-per-plan”:“1000”,“node-id”:“node01”,“wal-replay-concurrency-limit”:“1”,“data-dir”:“/var/lib/influxdb3/data”,“log-filter”:“info,influxdb3_enterprise=debug,influxdb3_compactor=debug”,“wal-snapshot-size”:“300”,“compaction-check-interval”:“30m”,“num-io-threads”:“2”,“force-snapshot-mem-threshold”:“256”} |

  • docker inspect influxdb3-enterprise
    see attached file …

  • compose.yaml
    see attached file …

  • current docker_logs_influxdb3-enterprise.txt
    see attached file …

I think it can only be a small thing why the compactor is bypassing my old imported data!
Does anyone have any ideas?
Many greetings from Vienna, HG
PS: I could not upload files, nor allowed for new user ?!?

If you’re using a trial license, it may help to connect a dedicated compactor node to your cluster so that you can ensure the CPU is working on that task and that task alone.

But perhaps more relevantly, it looks like there’s an existing GitHub issue that’s very plausibly related to this, where other users are reporting similar issues that seem similar to yours, where work is ongoing: [v3] InfluxDB3 imported data with plugin - compaction not working · Issue #27301 · influxdata/influxdb · GitHub

I’d keep an eye on that, as whatever resolution pops up there will likely also be your resolution.


Hi, thanks, I also left a short message there saying I’m having the same problem!

I don’t really understand it; surely there must be many people switching from InfluxDB2 who want to import their old data?!

Best regards, Harald