-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
Add retry on image push 5xx errors #2043
Conversation
Some registries can be flaky and return intermittent 5xx errors. This change allows those errors to be retried, similarly to network-level errors. Note that this needs the upstream containerd fix containerd/containerd#5276 to work reliably. This was tested with a registry that was modified to return 504 on every other manifest PUT. Without the change, exports to the registry fail every other attempt. With the change and the related containerd change, exports to the registry always succeed. Signed-off-by: Aaron Lehmann <alehmann@netflix.com>
@@ -65,7 +65,7 @@ func CopyChain(ctx context.Context, ingester content.Ingester, provider content. | |||
handlers := []images.Handler{ | |||
images.ChildrenHandler(provider), | |||
filterHandler, | |||
retryhandler.New(remotes.FetchHandler(ingester, &localFetcher{provider}), nil), | |||
retryhandler.New(remotes.FetchHandler(ingester, &localFetcher{provider}), func(_ []byte) {}), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Passing nil
to retryhandler.New
would cause panics if the retry handler tried to log something.
@@ -133,6 +133,7 @@ func (sw *streamWriter) Close() error { | |||
func LoggerFromContext(ctx context.Context) func([]byte) { | |||
return func(dt []byte) { | |||
pw, _, _ := progress.FromContext(ctx) | |||
defer pw.Close() |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I found that the client was getting "context cancelled" errors reading status if and only if the retry handler logged anything. It was a bit hard to track down the cause, but it looks like this missing Close
kept the status stream open after the build was complete and led to a timeout.
Some registries can be flaky and return intermittent 5xx errors. This
change allows those errors to be retried, similarly to network-level
errors.
Note that this needs the upstream containerd fix
containerd/containerd#5276 to work reliably.
This was tested with a registry that was modified to return 504 on every
other manifest PUT. Without the change, exports to the registry fail
every other attempt. With the change and the related containerd change,
exports to the registry always succeed.
Signed-off-by: Aaron Lehmann alehmann@netflix.com