fix: reset task status to queued on intermediate BullMQ retry attempts

[DaxServer]

When a job fails with a retryable error (SourceCdnError, HashLockError) and BullMQ schedules a retry, the failed event handler was returning early without updating the DB. Since status was set to in_progress at job start (line 122), it stayed there until the next attempt ran — making tasks appear stuck.

Fix: call updateUploadStatus(id, 'queued') in the early-return path before BullMQ retries. The access token is not touched, so retries continue to work.

— Claude Sonnet 4.6

[greptile-apps]

Confidence Score: 5/5

Safe to merge — the change is narrowly scoped to the intermediate-retry branch, is wrapped in a try/catch, and is covered by new and updated tests.

The fix is small and self-contained: one new DB call in a guarded try/catch, the access-token logic is untouched, and both the happy path and the DB-error path are exercised by the updated test suite.

No files require special attention.

Important Files Changed

Filename	Overview
backend/src/workers/upload.worker.ts	Adds updateUploadStatus(uploadId, 'queued') inside a try-catch in the intermediate-retry path of the failed event handler, so tasks no longer appear stuck at in_progress between BullMQ retry attempts. Comment updated to match new behavior.
backend/src/tests/upload.worker.test.ts	Updates existing intermediate-retry test to assert updateUploadStatus is called with 'queued', and adds a new test proving DB errors are swallowed so the handler never rejects.

Sequence Diagram

%%{init: {'theme': 'neutral'}}%%
sequenceDiagram
    participant BullMQ
    participant FailedHandler as failed event handler
    participant DB as uploads DB
    participant Logger

    BullMQ->>FailedHandler: emit('failed', job, err) [intermediate attempt]
    FailedHandler->>FailedHandler: "check !(err instanceof StorageError) && attemptsMade < attempts"
    FailedHandler->>Logger: warn — job attempt failed, will retry
    FailedHandler->>DB: updateUploadStatus(uploadId, 'queued')
    alt DB write succeeds
        DB-->>FailedHandler: ok
    else DB write fails
        DB-->>FailedHandler: throws
        FailedHandler->>Logger: error — failed to update db status
    end
    FailedHandler-->>BullMQ: return (early)
    BullMQ->>BullMQ: schedule next retry attempt

    Note over BullMQ,Logger: Final failure path (unchanged)
    BullMQ->>FailedHandler: emit('failed', job, err) [last attempt]
    FailedHandler->>DB: updateUploadStatus(uploadId, 'failed') + clearUploadAccessToken

Loading

%%{init: {'theme': 'base', 'themeVariables': {"darkMode": true, "background": "#0d1117", "primaryColor": "#21262d", "primaryTextColor": "#e6edf3", "primaryBorderColor": "#8b949e", "lineColor": "#8b949e", "textColor": "#e6edf3", "edgeLabelBackground": "#161b22", "actorBkg": "#21262d", "actorBorder": "#8b949e", "actorTextColor": "#e6edf3", "actorLineColor": "#8b949e", "signalColor": "#8b949e", "signalTextColor": "#e6edf3", "noteBkgColor": "#373320", "noteBorderColor": "#d4a72c", "noteTextColor": "#f0e6c0", "labelBoxBkgColor": "#21262d", "labelBoxBorderColor": "#8b949e", "labelTextColor": "#e6edf3", "loopTextColor": "#e6edf3", "activationBkgColor": "#30363d", "activationBorderColor": "#8b949e"}}}%%
sequenceDiagram
    participant BullMQ
    participant FailedHandler as failed event handler
    participant DB as uploads DB
    participant Logger

    BullMQ->>FailedHandler: emit('failed', job, err) [intermediate attempt]
    FailedHandler->>FailedHandler: "check !(err instanceof StorageError) && attemptsMade < attempts"
    FailedHandler->>Logger: warn — job attempt failed, will retry
    FailedHandler->>DB: updateUploadStatus(uploadId, 'queued')
    alt DB write succeeds
        DB-->>FailedHandler: ok
    else DB write fails
        DB-->>FailedHandler: throws
        FailedHandler->>Logger: error — failed to update db status
    end
    FailedHandler-->>BullMQ: return (early)
    BullMQ->>BullMQ: schedule next retry attempt

    Note over BullMQ,Logger: Final failure path (unchanged)
    BullMQ->>FailedHandler: emit('failed', job, err) [last attempt]
    FailedHandler->>DB: updateUploadStatus(uploadId, 'failed') + clearUploadAccessToken

Loading

_{Reviews (2): Last reviewed commit: "fix: guard queued status reset in failed..." | Re-trigger Greptile}

[DaxServer]

Both fixed in 313ef39.

1. Wrapped the updateUploadStatus('queued') call in its own try-catch matching the existing error-handling pattern in the handler — a DB error now logs and the handler resolves cleanly.
2. Updated the comment from "skip DB updates" to "reset status to queued" to match the new behavior.

— Claude Sonnet 4.6