-
Notifications
You must be signed in to change notification settings - Fork 618
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 IPFS container start routine #3181
Conversation
73ea2cb
to
3e44585
Compare
Signed-off-by: apostasie <spam_blackhole@farcloser.world>
|
||
name := testutil.Identifier(base.T) | ||
// listen on 0.0.0.0 to enable 127.0.0.1 | ||
listenIP := net.ParseIP("0.0.0.0") |
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.
Why not 127.0.0.1?
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.
Then it will not allow access using (non-localhost) host-ip, reducing possible test scope (comment is slightly misleading, originating from previous code in the registry start function).
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.
Thanks
Thanks a lot @ktock |
TestIPFSAddress and TestIPFSComposeUp have been failing regularly on the CI (with a rather long timeout).
Furthermore, the initialization code was using static ports and container name, making it impossible to parallelize.
This PR rewrites the initialization code of the IPFS registry container after the regular registry container start method - giving guarantees that there is no port or name collision, a sensible timeout in case of failure, an actual test to verify the registry is up and running.
It also removes the unnecessary testing of
detached-ns
(apparently triggering failures with rootlesskit 1).CI is green on push (except for 1 unrelated failure involving systemctl that we have seen somewhere else).
Along with the other changes from today, this PR should increase CI reliability quite a bit.
PTAL at your convenience @AkihiroSuda @ktock .
Of course would love to see this in ASAP as failing CI has been a massive PITA, but lmk evidently if there is anything you would like tweaked.
Cheers.