Closed w-yue closed 2 years ago
It seems to be hung in kind-e2e-test. I need to debug my failure as I cannot get local repro. I'll quit this CI and restart later.
Found and fixed the issue that was causing CI failure in my PR. I've now restarted your CI.
Please see if we need to fix the eth0 reference in scripts/03_start.sh
Close this PR. Will resubmit after CI issue investigation.
https://github.com/CentaurusInfra/mizar/pull/555 is the new PR for the same issue.
What does this PR do? We currently uses hard-coded "eth0" as nic interface in our code which prevents Mizar from running in system whose default interface is not named as eth0, which is the case in later Ubuntu version or other Linux flavors.
This PR remove hard coded "eth0" from the logic and replaces with the default interface name obtained from the system. It is a fix for issue #519.
How was this tested? It is tested in an AWS 2 nodes cluster environment with default interface name as "ensxx", the k8s cluster was launched successfully and the default interface name is detected and used in Mizar pods(instead of eth0), the testpods were created successfully and ping-able to each other.
Are there any user facing / API changes? No.