Workload downloading stuck

[dim river]

As of recently, Salad completely stopped working for days now. The job downloading works fine at first then it reaches a point (beyond 10GB) and completely stalls with no network/disk activity. This re-occurs and replicates every single time the app is started or even completely re-installed with no way to resolve it. It started happening about 2 days ago and there were no changes done to the machine running Salad at all.

2025-08-06 06:01:59.970 +01:00 [INF] Workload Instance States: 
    ndm: { "workloadId": "ndm", "stopped": { "reason": "WORKLOAD_INSTANCE_STOPPED_REASON_EXITED", "runningDuration": "58.846028100s", "startedAt": "2025-08-06T03:36:35.238883300Z", "stoppedAt": "2025-08-06T03:37:34.084911400Z" } },
    8318309d-b622-4db2-b6c6-95081ada90ce: { "workloadId": "8318309d-b622-4db2-b6c6-95081ada90ce", "stopped": { "reason": "WORKLOAD_INSTANCE_STOPPED_REASON_SERVER_REQUESTED", "reasonDetail": "MatrixUnassign:InstallCancelled", "runningDuration": "0s" } },
    systeminformation: { "workloadId": "systeminformation", "stopped": { "reason": "WORKLOAD_INSTANCE_STOPPED_REASON_EXITED", "runningDuration": "5.610321200s", "startedAt": "2025-08-06T03:36:35.150445400Z", "stoppedAt": "2025-08-06T03:36:40.760766600Z" } }
2025-08-06 06:01:59.971 +01:00 [INF] Node Image Cache States: 
    { "imageRef": "registry-eu-01.salad.com/cache-images/5524535a-fd79-45cd-aa98-db9da8cc51bc:latest", "status": "CONTAINER_IMAGE_STATUS_PULLING", "ttl": "2591999.793763115s", "startedAt": "2025-08-06T03:37:44.311351951Z", "stoppedAt": "0001-01-01T00:00:00Z", "pullDuration": "-9223372036.854775808s" }

[plush bridge]

#official-support and include your logs

[stiff dawn]

One of my 9 devices has been doing the same thing all day. It doesn't matter what I do it's stuck on downloading a job and doesn't progress. Did you figure out your problem?

[plush bridge]

It's still an issue, unknown cause afaik.

temporary fix is to restart the device

[warm prism]

I had the problem again with downloading the job, running and then, after around 10 seconds, finding the job again.