No idea when or how this happened, but it's a weird one... While poking around our production system DB today, I noticed there was a history table from 1986. Only one entry in it, for a tag that wasn't created until 2021 according the to the sqlth_te
table.
MariaDB [ignitiondb]> SELECT * FROM sqlt_data_1_1986_12;
+-------+----------+------------+----------------------------------+-----------+---------------+--------------+
| tagid | intvalue | floatvalue | stringvalue | datevalue | dataintegrity | t_stamp |
+-------+----------+------------+----------------------------------+-----------+---------------+--------------+
| 60795 | NULL | NULL | Modem eDRX value req: 10.24 sec | NULL | 192 | 534714917336 |
+-------+----------+------------+----------------------------------+-----------+---------------+--------------+
1 row in set (0.030 sec)
MariaDB [ignitiondb]> SELECT * FROM sqlth_te WHERE id = 60795;
+-------+-----------------------------------------+------+----------+-----------+---------------+---------+
| id | tagpath | scid | datatype | querymode | created | retired |
+-------+-----------------------------------------+------+----------+-----------+---------------+---------+
| 60795 | edge nodes/none/foo/console | 2 | 2 | 0 | 1632253751878 | NULL |
+-------+-----------------------------------------+------+----------+-----------+---------------+---------+
1 row in set (0.033 sec)
This isn't really breaking anything, and isn't really a problem unless it happens again, but WTF?! That shouldn't happen even if the tag did somehow get a reading with a timestamp back then, since it should've made a new sqlth_te
entry with a created date back then. This entry is outside the created/retired range for the entry it used.
I have no expectation of a fix, and I'm not even really complaining. I just felt like sharing this moment of weirdness with an audience that would understand.