Hi @<1689446563463565312:profile|SmallTurkey79> ! I will take a look at this and try to replicate the issue. In the meantime, I suggest you look into other dependencies you are using. Maybe some dependency got upgraded and the upgrade now triggers this behaviour in clearml.
odd bc I thought I was controlling this... maybe I'm wrong and the env is mis-set.
yeah, it just shows what I see in the Console, but then immediately goes back to polling for more work (so... instead of running backtest, it exits, no completion message)
damn. I can't believe it. It disappeared again despite having 1.15.1 be the task's clearml version.
I'm going to try running the pipeline locally.
do you have any STATUS REASON under the INFO section of the controller task?
worker thinks its in venv mode but is containerized .
apiserver is docker compose stack
ill check logs next time i see it .
currently rushing to ship a model out, so I've just been running smaller experiments slowly hoping to avoid the situation . fingers crossed .
damn, it just happened again... "queued" steps in the viz are actually complete. the pipeline task disappeared again without completion, logs mid-stream.
trying to run the experiment that kept failing right now, watching logs (they go by fast)... will try to spot anything anamolous
I have tried other queues, they're all running the same container.
so far the only thing reliable is pipe.start_locally()
yeah this problem seems to happen on 1.15.1 and 1.16.2 as well, prior runs were on the same version even. It just feels like it happens absolutely randomly (but often).
just happened again to me.
The pipeline is constructed from tasks, it basically does map/reduce. prepare data -> model training + evaluation -> backtesting performance summary.
It figures out how wide to go by parsing the date range supplied as input parameter. Been running stuff like this for months but only recently did things just start... vanishing like this.
Would appreciate any help. Really need this to be more robust to make the case for company-wide adoption.
(the "magic" of the env detection is nice but man... it has its surprises)
I think i've narrowed this down to the ssh connection approach.
regarding the container that runs the pipeline:
- when I made it stop using autossh tunnels and instead put it on the same machine as the clearml server + used docker network host mode, suddenly the problematic pipeline started completing.
it's just so odd that the pipeline controller task is the only one with an issue. the modeling / data-creation tasks really all seem to complete consistently just fine.
so yeah, best guess now is that its unrelated to clearml verison but rather to the connectivity of the pipeline controller task to the api server.
when I run this pipeline controller locally (also using the same ssh tunnel approach for comms), the pipeline completes just fine. so it's something specific about how its working inside the container vs on my machine, it seems.
when i run the pipe locally, im using the same connect.sh script as the workers are in order to poll the apiserver via the ssh tunnel.
did you take a look at my connect.sh script? I dont think it's a problem since only the controller task is the problem.
Is there some sort of culling procedure that kills tasks by any chance? the lack of logs makes me think it's something like that.
I can also try different agent versions.
ah. a clue! it came right below that but i guess out of order...
that id is the pipeline that failed
it happens consistently with this one task that really should be all cache.
I disabled cache in the final step and it seems to run now.
are you running this locally or are you enqueueing the task (controller)?
N/A (still shows as running despite Abort being sent)
yeah locally it did run. I then ran another via UI spawned from the successful one, it showed cached steps and then refused to run the bottom one, disappearing again. No status message, no status reason. (not running... actually dead)
Hi @<1689446563463565312:profile|SmallTurkey79> !Prior runs of this pipeline worked just fine What SDK version were you using for the prior runs? Does this still happen if you revert to that version?
Can you provide a script that imitates what you are doing?
In the pipeline you are running, are you creating new tasks/pipelines/datasets?
the workers connect to the clearml server via ssh-tunnels, so they all talk to "localhost" despite being deployed in different places. each task creates artifacts and metrics that are used downstream
would it be on the pipeline task itself then, since that's what's disappearing?
I will do some experiment comparisons and see if there are package diffs. thanks for the tip.
None here's how I'm establishing worker-server (and client-server) comms fwiw
would it be on the pipeline task itself then, since that's what's disappearing? that likely the case