hubble/relay: Fix certificate reloading in PeerManager #31376
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
d0f9ea8 to
a4a2d10
Compare
glrf
commented
Mar 14, 2024
a4a2d10 to
b98794b
Compare
chancez
reviewed
Mar 14, 2024
chancez
approved these changes
Mar 14, 2024
b98794b to
143d5a5
Compare
|
/test |
This change fixes an issue with the PeerManager that can lead to Relay being unable to reconnect to some or all peers when the client certificate expires or the Certificate Authority is replaced. Before this change, when the client certificate changes, we did not redial or update the exiting gRPC ClientConns. When the old certificate becomes invalid, (expiring, changed CA, or revoked) The connection will eventually fail with a certificate error. However, the gRPC ClientConn is not closed, but treats the certificate error as a transient failure and will retry connecting with the old credentials indefinitely. In most cases this will cause the relay health checks to fail. Relay will restart and successfully reconnect to all peers. However, if a new peer joins between the certificate being updated and the connections failing, Relay may keep on running in a degraded state. This issue was introduced by cilium#28595. Before that change, Relay aggressively closed and re-dialed ClientConns on any error, mitigating this problem. We fix this issue by wrapping the provided gRPC transport credentials and updating the TLS configuration whenever a new TLS connection is established. This means every TLS connection will use up-to-date certificates and gRPC ClientConns will be able to recover when their certificate changes. Fixes: aca4d42 ("hubble/relay: Fix connection leak when reconnecting to peer service") Signed-off-by: Fabian Fischer <fabian.fischer@isovalent.com>
|
Hit flake that should be fixed by #31381 Rebasing on main |
143d5a5 to
b2035f6
Compare
|
/test |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This change fixes an issue with the PeerManager that can lead to Relay being unable to reconnect to some or all peers when the client certificate expires or the Certificate Authority is replaced.
Before this change, when the client certificate changes, we did not redial or update the exiting gRPC ClientConns. When the old certificate becomes invalid, (expiring, changed CA, or revoked) The connection will eventually fail with a certificate error.
However, the gRPC ClientConn is not closed, but treats the certificate error as a transient failure and will retry connecting with the old credentials indefinitely.
In most cases this will cause the relay health checks to fail. Relay will restart and successfully reconnect to all peers. However, if a new peer joins between the certificate being updated and the connections failing, Relay may keep on running in a degraded state.
This issue was introduced by #28595. Before that change, Relay aggressively closed and re-dialed ClientConns on any error, mitigating this problem.
We fix this issue by wrapping the provided gRPC transport credentials and updating the TLS configuration whenever a new TLS connection is established. This means every TLS connection will use up-to-date certificates and gRPC ClientConns will be able to recover when their certificate changes.
Fixes: aca4d42 ("hubble/relay: Fix connection leak when reconnecting to peer service")
Please ensure your pull request adheres to the following guidelines:
description and a
Fixes: #XXXline if the commit addresses a particularGitHub issue.