Failed to save: uploadChunk failed: Cache service responded with 500

[itss050]

Looking for help on how to debug cache failing to save.

I have two separate instances of GHES running. One works as expected however the other fails saving the cache. I am unable to find anything in the logs and the workflow output isn't helpful.

##[debug]File Size: 3395 ##[debug]Reserving Cache ##[debug]Resource Url: https://mydomain.com/_services/artifactcache/U0Hz6J5v04j8ziLeRZlN5c5BcCqyGoyz6VhmnrktxpMkvCN6JH/_apis/artifactcache/caches ##[debug]Saving Cache (ID: 108) ##[debug]Upload cache ##[debug]Resource Url: https://mydomain.com/_services/artifactcache/U0Hz6J5v04j8ziLeRZlN5c5BcCqyGoyz6VhmnrktxpMkvCN6JH/_apis/artifactcache/caches/108 ##[debug]Upload concurrency: 4 ##[debug]Upload chunk size: 33554432 ##[debug]Awaiting all uploads ##[debug]Uploading chunk of size 3395 bytes at offset 0 with content range: bytes 0-3394/* ##[debug]uploadChunk (start: 0, end: 3394) - Attempt 1 of 2 failed with error: Cache service responded with 500 ##[debug]uploadChunk (start: 0, end: 3394) - Error is not retryable Warning: Failed to save: uploadChunk (start: 0, end: 3394) failed: Cache service responded with 500 Cache saved with key: Linux-primes-v5 ##[debug]Node Action run completed with exit code 0 ##[debug]Finishing: Save Primes

The UI shows conflicting info on a cache existing. Calling APIs directory returns no results in the cache

[Tehmeena0]

It sounds like a frustrating issue, especially with the logs not providing much clarity. Since one of your GHES instances works fine while the other doesn't, it might be worth checking if there are any differences in the configuration or network settings between the two. Also, ensure that both instances are running the same versions of GHES and that there's no proxy or firewall interference on the failing instance.

If possible, try increasing the logging level to see if it provides more detailed information. Sometimes, even a small discrepancy can cause issues like this. If the problem persists, you might want to reach out to GitHub Support for further investigation.