latentspace
RRunPod
•Created by latentspace on 10/30/2024 in #⚡｜serverless
Request stuck because of exponential backoff, what does it mean?
relevant log line:
"message":"b2http.py :600 2024-10-29 15:56:21,141 Pausing thread for 64 seconds because that is what the default exponential backoff is\n"
this keeps repeating in serverless GPU logs and pod is on and request is stuck
6 replies
RRunPod
•Created by latentspace on 10/30/2024 in #⚡｜serverless
in serverless CPU, after upgrading to runpod sdk 1.7.4, getting lots of "kill worker" error.
This is severless CPU workers, not GPU. Initially serverless CPU was on 1.7.3 and was timing out if execution time was longer than a minute. So, I downgraded to 1.6.2 and it worked fine. Yesterday, I upgraded to 1.7.4 and getting "kill worker" error
1 replies