Hey!
I don't know how to reproduce this yet, this happens randomly.
I have a simple DAG:
name: test-dag
schedule: "0/2 * * * *"
steps:
- echo "Step 1 - Starting workflow"
- echo "Step 2 - Processing data"
- echo "Step 3 - Workflow complete"
Dagu is run using this compose file:
x-shared-parameters: &shared-parameters
image: ghcr.io/dagu-org/dagu:2.3.4
entrypoint: dagu
environment:
DAGU_HOME: /home/dagu/data
DAGU_DAGS_DIR: /home/dagu/config/dags
DAGU_BASE_CONFIG: /home/dagu/config/base.yaml"
DAGU_DATA_DIR: /home/dagu/data
DAGU_LOG_DIR: /home/dagu/logs
volumes:
- /var/run/docker.sock:/var/run/docker.sock
- ./data:/home/dagu/data
- ./config/config.yaml:/home/dagu/data/config.yaml
- ./config:/home/dagu/config
- ./logs:/home/dagu/logs
services:
ui:
<<: *shared-parameters
ports:
- 8080:8080
command: ["server"]
scheduler:
<<: *shared-parameters
deploy:
mode: replicated
replicas: 3
command: ["scheduler"]
coordinator:
<<: *shared-parameters
depends_on:
- scheduler
deploy:
mode: replicated
replicas: 3
command: ["coordinator"]
worker:
<<: *shared-parameters
depends_on:
- coordinator
- scheduler
deploy:
mode: replicated
replicas: 2
command: ["worker"]
Sometimes the task gets stuck at Queued state and stays there:
This goes away only if I delete the run in question from /home/dagu/data/dag-runs/test-dag/dag-runs/2026/03/23/dag-run_20260323_091200Z_019d19f7-2302-7686-8499-fc839bee9648
Logs do not appear for this run so I cannot attach those. It would be nice the scheduler could get this "unstuck" somehow by running it again after some time. In the meantime, how do I force-start such stuck tasks?
Hey!
I don't know how to reproduce this yet, this happens randomly.
I have a simple DAG:
Dagu is run using this compose file:
Sometimes the task gets stuck at
Queuedstate and stays there:This goes away only if I delete the run in question from
/home/dagu/data/dag-runs/test-dag/dag-runs/2026/03/23/dag-run_20260323_091200Z_019d19f7-2302-7686-8499-fc839bee9648Logs do not appear for this run so I cannot attach those. It would be nice the scheduler could get this "unstuck" somehow by running it again after some time. In the meantime, how do I force-start such stuck tasks?