Like with my comments in #38, I noticed that networking was initially dead - at least the first time I ran it. I was not running any special setup commands to install or remove any packages in the test bed, other than whatever the hello package is asking the package manager to pull.
I would have attached my aptest.log again but for good measure I wanted to replicate the same issue. So I ran the command again, and again, and again, and it seems to work now clobbering the old log. Let's hope it stays that way and I'm not simply catching a race condition when it's in a good mood.
Hello Dominique,
I prepared an autopkgtest container like so with the 6.11-rc4 kernel installed:
$ autopkgtest- buildvm- ubuntu- cloud --verbose -r noble -a amd64 --cloud-image-url http:// cloud-images. ubuntu. com/daily/ server --post-command "wget -c https:/ /kernel. ubuntu. com/mainline/ v6.11-rc4/ amd64/linux- headers- 6.11.0- 061100rc4- generic_ 6.11.0- 061100rc4. 202408182119_ amd64.deb https:/ /kernel. ubuntu. com/mainline/ v6.11-rc4/ amd64/linux- headers- 6.11.0- 061100rc4_ 6.11.0- 061100rc4. 202408182119_ all.deb https:/ /kernel. ubuntu. com/mainline/ v6.11-rc4/ amd64/linux- image-unsigned- 6.11.0- 061100rc4- generic_ 6.11.0- 061100rc4. 202408182119_ amd64.deb https:/ /kernel. ubuntu. com/mainline/ v6.11-rc4/ amd64/linux- modules- 6.11.0- 061100rc4- generic_ 6.11.0- 061100rc4. 202408182119_ amd64.deb && sudo dpkg -i *.deb"
That builds a container image and then downloads the mainline kernel with, unless I'm mistaken, the patch in there from:
https:/ /kernel. ubuntu. com/mainline/ v6.11-rc4/
I then attempt to perform CI testing for the GNU hello package with the fresh test bed:
$ sudo TMPDIR=$PWD autopkgtest --apt-upgrade -ddd -U ./hello_ 2.10-3build1. dsc -- qemu --qemu- options= -enable- kvm --cpus=8 --ram-size=8192 --debug --show-boot autopkgtest- noble-amd64. img |& ts %.T | /usr/bin/tee ./aptest.log
Like with my comments in #38, I noticed that networking was initially dead - at least the first time I ran it. I was not running any special setup commands to install or remove any packages in the test bed, other than whatever the hello package is asking the package manager to pull.
I would have attached my aptest.log again but for good measure I wanted to replicate the same issue. So I ran the command again, and again, and again, and it seems to work now clobbering the old log. Let's hope it stays that way and I'm not simply catching a race condition when it's in a good mood.