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
test: Retrieve the private interface in an Eventually #16990
Conversation
Upon sysdump inspection of cilium#16479, it's possible that the failures in retrieving the private interface are time-sensative. As cilium#16479 (comment) points out, the output of `ip a` actually showed that the interface the test was looking for existed, so it's very likely a timing issue. This commit attempts to fix this flake by retrying until we are able to extract out the interface. Signed-off-by: Chris Tarazi <chris@isovalent.com>
417a46b
to
47dad6f
Compare
test-1.16-netnext |
test-1.20-4.19 |
test-1.19-5.4 |
Do we have any idea why this would be happening? When the test fails, it's usually not the first test to run, so the enp0s8 interface has been set up for a while already. |
@pchaigno It might have something to do with Cilium being re-installed and it re-configuring the devices. I tried reproducing locally but I couldn't even when forcing the relevant previous test to run as #16479 (comment) pointed out. I can't think of any other explanation because the test actually spits out Actually another thing that I just thought about is that it could be netlink being unreliable. It's possible that the first call to |
That's what I had in mind. Ideally, we'd print, in the error message, the exact
Yep, you're probably right. |
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.
Does this PR really need a review from @cilium/bpf? |
No, but we can't remove the team review request, so Chris removed the reviewer. I guess it can be merged now if tests are passing. |
I'll let Chris mark as |
test-1.19-5.4 |
test-1.19-5.4 |
Upon sysdump inspection of
#16479, it's possible that the
failures in retrieving the private interface are time-sensative. As
#16479 (comment)
points out, the output of
ip a
actually showed that the interface thetest was looking for existed, so it's very likely a timing issue. This
commit attempts to fix this flake by retrying until we are able to
extract out the interface.
Fixes: #16479