Navigation Menu

Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix/enable body read retries and logging #194

Merged
merged 3 commits into from Feb 3, 2019

Conversation

JohnRusk
Copy link
Member

@JohnRusk JohnRusk commented Feb 3, 2019

Two related changes to the way we read response bodies in downloads:

  1. Log when retries happen
  2. Force a retry when a chunk looks like its taking too long. We do this by closing the reader from a different Go routine, which renders it unusable on the "main" go routine. We do this because otherwise we occasionally get chunks which are so slow that they hold up our sequential writing to a significant degree. Forced retrying allows them to complete faster.

Without this, we can see very occasional chunks that are orders of magnitude slower than normal.  That becomes a problem given our desire to process the files sequentially (for MD5 computation and, in some cases, disk performance). Root cause of the slowness is yet to be determined, but this workaround is effective for now.
@JohnRusk JohnRusk merged commit ddf4ed2 into dev Feb 3, 2019
@JohnRusk JohnRusk deleted the fix/enable-body-read-retries-and-logging branch February 8, 2019 02:30
rickle-msft pushed a commit to rickle-msft/azure-storage-azcopy that referenced this pull request Feb 19, 2019
* Port forced retry and better logging to the azbfs retry reader

(Same as already in Blob/File SDKs)

* Enable logging of retries in response body downloads

* Force body read retries when a downloaded chunk is abnormally slow

Without this, we can see very occasional chunks that are more than an order of magnitude slower than normal.  That becomes a problem given our desire to process the files sequentially (for MD5 computation and, in some cases, disk performance). Root cause of the slowness is yet to be determined, but this workaround is effective for now.

* Update "dep" dependency of blob SDK to latest version
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

1 participant