While Server Update Running Schedule task is not updating

[milansoliya4210]

To Reproduce

Bug Report: Scheduled Task Continues Running After Server Update While Dokploy Shows It as Cancelled

Description

I found an issue related to scheduled tasks during a Dokploy server update.

When a Dokploy server update is performed while a scheduled task (Cron/Scheduled Job) is currently running, the task continues executing in the background even after the server comes back online. However, Dokploy incorrectly reports that the scheduled task was cancelled, and the service status becomes Stopped, requiring a manual redeploy.

This creates a mismatch between the actual runtime state and the state shown in the Dokploy UI.

Steps to Reproduce

1. Deploy a service that contains a scheduled task (Cron Job).
2. Start a scheduled task that runs for a noticeable amount of time.
3. While the scheduled task is still running, perform a Dokploy server update.
4. Wait for the server to restart and become available again.
5. Open the service in the Dokploy dashboard.

Expected Behavior

• Running scheduled tasks should either:

  • be gracefully terminated during the update, or
  • be properly tracked after the server restarts.

• The Dokploy UI should accurately reflect the actual state of the scheduled task.

• The service should remain in a consistent state without requiring manual intervention.

Actual Behavior

• The scheduled task continues running in the background after the server update.
• Dokploy marks the scheduled task as Cancelled.
• The service status is shown as Stopped.
• The UI no longer reflects the actual running background process.
• The service must be manually redeployed to restore the correct state.

Impact

This can lead to:

• Incorrect monitoring information.
• Confusion about whether scheduled jobs are still executing.
• Background processes continuing without being visible in the UI.
• Risk of duplicate executions if the user restarts or redeploys the service, assuming the original job has stopped.

Environment

• Dokploy Version: (please fill in)
• Docker Version: (please fill in)
• Operating System: (please fill in)
• Deployment Type: (VPS/Dedicated Server/etc.)

Additional Information

It appears that after the Dokploy server update, the scheduler or process state is not synchronized with the actual running container/process. The UI reports the scheduled task as cancelled even though the underlying process continues to execute in the background.

This seems to be a state synchronization issue between Dokploy and the running scheduled task after the server update.

Current vs. Expected behavior

Following the steps described above, I expected Dokploy to either gracefully stop the running scheduled task during the server update or correctly track its state after the server restarted.

Instead, I observed that:

• The scheduled task continues running in the background even after the Dokploy server update is completed.
• In the Dokploy UI, the scheduled task is marked as Cancelled, even though the process is still actively running.
• The service status is shown as Stopped, despite the background task continuing to execute.
• To restore the service to a normal state, I have to manually redeploy it.

This results in an inconsistent state between the actual running process and what Dokploy reports in the UI. It can also lead to confusion and potentially duplicate job executions if the service is redeployed while the original scheduled task is still running.

Provide environment information

Operating System:
OS: Ubuntu 20.04
Arch: arm64

Dokploy version:
0.29.14

VPS Provider:
Hostinger

What applications/services are you trying to deploy?
Docker Compose application with a scheduled (Cron) task for python image.

Which area(s) are affected? (Select all that apply)

Application

Are you deploying the applications where Dokploy is installed or on a remote server?

Same server where Dokploy is installed

Additional context

No response

Will you send a PR to fix it?

No

[linear-code]

DOK-613

[dosubot]

Hi @milansoliya4210! I'm Dosu and I'm helping the Dokploy team.

I can confirm this is a real state synchronization gap. Here's what's happening under the hood:

On every server startup, initCancelDeployments() runs a blanket update that marks all deployments with status = "running" as "cancelled" — it does not distinguish between regular deployments and scheduled task executions. It then resets the related application/compose status to "idle" [1].

Separately, initSchedules() re-registers all enabled cron schedules with node-schedule, but it only sets up future triggers — it does not check whether any Docker containers from previous scheduled task executions are still running.

Since scheduled tasks for application/compose types work by running docker exec inside the existing service container [2], the container itself isn't stopped by Dokploy's restart. The docker exec process also continues independently of the Dokploy server process. So the sequence is:

1. Scheduled task fires → deployment record set to "running" → command runs via docker exec
2. Dokploy server restarts (update)
3. initCancelDeployments() marks the deployment as "cancelled"
4. initSchedules() re-registers future cron triggers but performs no reconciliation
5. The docker exec process finishes (or is still running), but Dokploy has already marked it cancelled

There is currently no reconciliation mechanism in the startup sequence that checks actual container/process state against deployment records. This is the root cause of the mismatch you're seeing.

A fix would likely need to either:

• Gracefully terminate in-progress docker exec processes before the server shuts down for an update, or
• Add post-restart reconciliation that checks whether containers tied to "running" scheduled task deployments are still active, and updates the deployment status accordingly instead of blanket-cancelling everything.

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.