Description of problem:
Currently oc-mirror v2 defaults related to concurrency are too high and some container registries are rejecting so many requests. Some customer are using some flags to avoid the network issues: In this conversation it is possible to see that --parallel-images 2 --retry-delay 2s --retry-times 5 are used https://redhat-internal.slack.com/archives/C02JW6VCYS1/p1749568315478749?thread_ts=1749566735.317039&cid=C02JW6VCYS1 there is also another customer using a similar set of flags: --retry-times 30 --image-timeout 60m --parallel-images 2 --retry-delay 2s
Version-Release number of selected component (if applicable):
How reproducible:
It is not always reproducible, but the dev team also faced this issue so it is a known issue.
Steps to Reproduce:
1. 2. 3.
Actual results:
Expected results:
Additional info:
Reducing the defaults could be a solution. Also changing the default of --retry-delay to zero would change the behavior inside of containers/common/pkg/retry from linear to exponential backoff (which could help when retrying pushing the image)
- blocks
-
OCPBUGS-58279 Reduce the default to avoid network issues
-
- Closed
-
- is cloned by
-
OCPBUGS-58279 Reduce the default to avoid network issues
-
- Closed
-
- links to