Is there a chance the BEFORE trigger functions are doing something that could be leading to the error?
In the error log is there a line with the actual values that failed?
The error log does not show the literal values, no. Here is a literal example from the logs:
2020-03-20 19:51:11 NZDT [15165]: [6-1] ERROR: duplicate key value violates unique constraint "_hyper_1_1931_chunk_da_datum_x_acc_idx"
2020-03-20 19:51:11 NZDT [15165]: [7-1] CONTEXT: SQL statement "INSERT INTO solardatum.da_datum(ts, node_id, source_id, posted, jdata_i, jdata_a, jdata_s, jdata_t)
VALUES (ts_crea, node, src, ts_post, jdata_json->'i', jdata_json->'a', jdata_json->'s', solarcommon.json_array_to_text_array(jdata_json->'t'))
ON CONFLICT (node_id, ts, source_id) DO UPDATE
SET jdata_i = EXCLUDED.jdata_i,
jdata_a = EXCLUDED.jdata_a,
jdata_s = EXCLUDED.jdata_s,
jdata_t = EXCLUDED.jdata_t,
posted = EXCLUDED.posted
RETURNING (xmax = 0)"
PL/pgSQL function solardatum.store_datum(timestamp with time zone,bigint,text,timestamp with time zone,text,boolean) line 10 at SQL statement
2020-03-20 19:51:11 NZDT [15165]: [8-1] STATEMENT: select * from solardatum.store_datum($1, $2, $3, $4, $5) as result
As for the BEFORE triggers, the solardatum.trigger_agg_stale_datum one does an INSERT into a different table and a SELECT from this same table. The _timescaledb_internal.insert_blocker one is part of the TimescaleDB extension which looks like it wouldn’t have an impact to this issue, but the source of that is
— m@
