I have followed setting up the timescaledb schema as in: New integration: TimescaleDB. The issue is when I try to connect using the latest version of the timescale backend, I get:
ERROR: relation “edge_channel” does not exist
Looking through the code it looks like the issue is at:
Cache fromDatabase(Statement stmnt) throws SQLException
Seems like I am querying table edge_channel even though it does not exist. I am a bit of noob when it comes to databases, so sorry if this is a stupid question. Would you be able to supply me with the latest configuration of TimescaleDB? From what I can see, there are a couple of fields that I am missing from the one that is described above.
Then lastly, are there any active development still happening on Timescaledb? Or is this not really priority at the moment?
Hi Openems folks,
Just an update:
I did some further digging and its clear that the schema has been updated from the original post here: New integration: TimescaleDB.
From what I can see, the issue is that the schema is not automatically created at start-up when the database is still empty, hence creating the issue above. So I guess the question is really what I should do different so that the schema is implemented if it does not yet exists? I manually tried to create the schema, similar to what I could find in
SchemaGenerator.java, but with no success.
Please would someone be able to point me in the right direction to getting my environment set up? I am eager to start working on the backfilling function of TimescaleDB.
Hope to hear from you soon.
sorry, I have been (and still am) currently very busy and hardly find time to answer in the Community and handle open Pull-Requests.
The TimescaleDB integration had been developed by our dev team at FENECON because we had planned to replace our existing InfluxDB 1.8 setup with it. Unfortunately after a few months of effort we had to admit, that the performance of TimescaleDB was just to slow for our requirements. Because of this we did not continue with TimescaleDB but instead built our own data-aggregation implementation based on a separate InfluxDB server.
Our setup now consists of two InfluxDB servers - one serving as a “data-warehouse” with values in high resolution - and the other one being optimized for queries from OpenEMS UI and only holding pre-aggregated data per-5-minutes or per-day. This setup is running in production on our side and we will hopefully soon publish it in OpenEMS.
We also have developed migration scripts that allow us to migrate from data-warehouse-influx to aggregated-influx. So if you now just use InfluxDB you will later be able to just add the ‘aggregated-influx’ and automatically migrate data.
So for now and until nobody else continues the work on TimescaleDB I would consider it deprecated and cannot recommend to use it with OpenEMS Backend.
Thanks for the feedback, it is much appreciated! And I am very excited to test it.
Do you have any idea on the timeline on the production version being available to the Openems community?
Hope to hear from you soon.