Open
Description
Feature Request
Which version of blobfuse was used?
2.1.0
Which OS distribution and version are you using?
Oracle Linux 8.8
If relevant, please share your mount command.
Using configuration file. See comment below.
What was the issue encountered?
Ref.: #1230
When exceeding the VM network bandwidth limit, we are getting "connection reset by peer" and "cancelled context".
Have you found a mitigation/solution?
azstorage: max-concurrency defaults to 32 threads (TCP connections) per file being handled.
This seem to max out the VM network bandwidth.
When manually "throttling" this to 8 or 16 we no longer seem to get errors.
Suggestion
We would like Blobfuse to somehow discover that it runs into this limitation and then automatically throttle the concurrency to avoid errors resulting in the client application (in this case Oracle RMAN backup) failing.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment