Closed ferricoxide closed 1 year ago
Note: the testing-environment sits behind a web proxy. Didn't (previously) think that it was a caching-proxy, but, right now, even though Packer is saying it's pulling fresh AMIgen8 content, what's executing says that it's actually pulling stale AMIgen8 content. I did a git show
of the PostBuild.sh
script against the remote that Packer says it's pulling fresh from: the git show
shows the correct content but the manner in which PostBuild.sh
is executing says that whatever Packer is pulling is stale. I think that the host from which Packer is executed transits the same proxy as the Packer build-VM, so, there shouldn't be any such differences, but, the supposition for the witnessed-behavior is all I got.
Going to try a new Packer run, tomorrow, and hope that things are a bit better behaved. If not, will need to further investigate possible sources for the seeming disconnect.
go codebuild go
Addresses need for RHEL8 – but not CentOS 8-stream or other relatives – builders per #394