Timescaledb compression is not deleting the original uncompressed chunk

, ,
1

Hi everyone, I’m currently using a Dockerized image of TimescaleDB version 2.15.0-pg15. We’ve set a chunk interval of 6 hours and configured a compression policy to run every 12 hours using the following command:

SELECT add_compression_policy(‘“deviceData”’, initial_start => NOW(), compress_created_before => INTERVAL ‘0s’);

However, after the compression process, we noticed that the size of the database did not change. Both uncompressed and compressed chunks remain, and the uncompressed chunks are not being deleted.

We have a retention policy of 30 days on the hypertable “deviceData,” along with two continuous aggregate refresh policies running every 1 minute and 5 minutes.

Could anyone advise on why the original uncompressed chunk is not getting automatically deleted after compression and how can we delete it? The current situation is not resolving our disk space issue, as it results in increased usage by retaining both chunk types.

image
image1514×1342 238 KB

2

@Divyanshu_Tripathi , chunks are not deleted after compression. The content of its is comressed into subchunks:

main_chunk
  |--compressed_chunk_for_main_chunk

QUERY PLAN                                                        |
------------------------------------------------------------------+
Append                                                            |
  ->  Custom Scan (DecompressChunk) on _hyper_1_1_chunk ht_1      |
        ->  Seq Scan on compress_hyper_2_2_chunk                  |