ERROR: XX000: tuple already updated by self

I created a hypertable, aggregation policy and retention policty. Everythink worked untill chunks were deleted (I am not sure if I they were deleted manually by me, or by retention policy), but now when I try to insert a row that would fall into deleted chunk I am getting an ERROR: tuple already updated by self.

Bug reproduction:

1. <hypertable with data>
2. `SELECT DROP_CHUNKS('data', newer_than => '2024-01-01 00:00:00.000+0200')`
3.  INSERT INTO data (time, value) VALUES ('2024-01-02 00:00:00.000+0200', 100)

TimescaleDB version: 2.14.2
PG version: 16.2
OS: Ubuntu 22.04.4
Installation method: Apt
Platform: Self-hosted

Do you have any idea what can I do?

Please, try with 2.15 that was released yesterday. If it does not work, would you mind to fill an issue on github?

I already created an issue: [Bug]: 'ERROR: tuple already updated by self' after inserting data when chunk was dropped · Issue #6904 · timescale/timescaledb · GitHub

We are using timescaledb in a production, about 100 000 000 rows are being added every day. We are hoping that new version is not going to break something else. BTW out case seems to be a common thing. It is weird that there are only like 2 issues about this

Thanks for reporting @DFFF. Glad that you have such high ingest rate :nerd_face: :rocket:

If you have more technical details about your setup or context, please, share directly in the github issue as the developers will have more data and insights about your usage to fix the case.

Thanks, we updated and now the problem is gone :slightly_smiling_face: