When uploading a job, the queue viewer says “checksum failure” and restarts the upload. Why?
This is caused by something on your computer modifying the compressed job archive between the time it was created and the time it finishes uploading. The compressed job archives live in the “outgoing” folder of the agent install directory until they are transferred. When the job archive is created, the agent on your computer sends a checksum of the file to our farm. If the checksum we generate on the received file is different from what your agent reports to us, then we must resend the job archive.
Real-time malware or virus scanning software is the most typical culprit here, which is why we strongly encourage you to exclude the entire agent install directory from real-time scanning. We’ve also seen very limited instances of some backup software modifying the file during a backup run. Lastly, it is possible (but not very likely) that this can be caused by failing memory on your machine. You would typically encounter a raft of other crashes/problems if that were the case. We only mention it here for the sake of completeness.
Once you configure the software on your system to insure nothing is modifying the compressed job archive during the upload, the checksums will match between both ends once transfer is complete. Matching checksums insures the transferred data is bit-for-bit accurate, and only such jobs will progress to the next phase of the rendering process.