Skip to content
Cindergrid

Troubleshooting

Updated 30 October 2026. Applies to scheduler 2026.11 and sync client 2.8.

Most failures fall into a handful of causes. If yours is not here, run cgsync job diagnose <job-id> and send the output to support@n71.assetlines.net.

Textures render black or pink

A path in the scene did not resolve. Run cgsync scene check and look under warnings. Common causes: a texture on a network drive that was offline at upload, a path with a trailing space, or an environment variable that exists only on your workstation.

Frame failed: out of memory

The job log shows a high-water mark and the node's limit. Move to a class with more RAM (C64 to C128, or C128 to M96 on CPU; G1 to G2 on GPU). For GPU renderers with out-of-core support, enable it before moving to a larger class: it is usually cheaper.

Uploads are slower than my line

  1. Run cgsync net test. It measures the path to the nearest region and shows how many streams help.
  2. Raise streams to your plan's maximum.
  3. Check for antivirus or backup software scanning the folder. Excluding the scene directory often doubles speed on Windows.
  4. Choose a region near you; transfers between regions are free if you later move.

Job is queued for a long time

Look at the queue panel on the status page. If the median wait is normal (a few minutes), your job is probably waiting on a limit: concurrency for your plan, free licence seats, or a dependency. The job page names which.

Exit code 6: hash mismatch

A downloaded chunk did not match the farm's hash, usually because of a flaky disk or a middlebox rewriting traffic. The client discards the chunk and fetches it again; if it recurs three times, the job pauses. Switch the network or set verify = "full" in the config.

Plan limit reached

Exit code 4 means storage, concurrency or monthly downloads hit your plan's cap. Usage above the allowance for downloads and storage is billed rather than blocked, unless you set a hard limit in account settings. Concurrency is a hard cap; upgrade or wait for running frames to finish.

Still stuck?

Studio and Pipeline customers can open a request from the job page and attach the diagnose bundle. Everyone else: contact page.