Hello all! Got an interesting situation happening here…recently we did a migration from 0.4.2 to 1.2.1 dagster. For the most part, things went smoothly. However, we are noticing that one of our jobs/schedules seems to be executing twice. Is there any way to track down what’s causing this phantom execution? Would restarting the daemon/dagit help? Any advice is appreciated! From what i can tell though, there’s only one schedule enabled. I do not see any residual schedules in dagit…
For context, this 17f8c99e run is the one appearing in the schedule’s history. this 0daa0479 run is seemingly being run by perhaps the “old” schedule? that is my guess as to what might be happening…
04/24/2023, 9:40 PM
Did I read correctly that you migrated from 0.4.2? Woah! A very early user
Regarding your issue, if it’s from that long ago the scheduler implementation completely changed - I wonder if somehow both the old and new scheduler are running? In any case - checking with the team
04/25/2023, 1:25 PM
haha yep, we’ve been using dagster since about august 2022.
I’m just trying to make sure there’s no remnants of an ‘old’ schedule left anywhere. I double checked some obvious places in the schema and it seems to have properly cleaned up older schedules. If there’s some DB schemas for postgres anywhere (quick search in dagster.io didn’t yield me any schemas/diagrams.) that’d be great!
but for what it’s worth, it seems that today’s scheduled runs properly triggered 1 run of this pipeline. this was after restarting our three services (dagit/daemon/grpc)