Hi
Penney:
We just installed on both Bare Metal and VM environments,
Confirm Result: Installs fine and when it reboots, it not goes back to the ISO boot menu.
the Install process for the following ISO, The verification result is the same.
ISO download path:
.
http://mirror.starlingx.cengn.ca/mirror/starlingx/master/centos/flock/20200520T015927Z/outputs/iso/bootimage.iso
VM Environment:
Host : Ubuntu 16.04 LTS 64-bit
VM Software : Virtual Machine Manager
Result:
Installs fine and when it reboots, it not goes back to the ISO boot menu.
Bare Metal Environment:
BIOS Setup:
USB disk is not set as the first boot item. (Manually select USB boot item)
Result:
Installs fine and when it reboots, it not goes back to the ISO boot menu.
----------------------------------------------
BIOS Setup:
USB disk is set as the first boot item.
Result:
Installs fine and when it reboots, it also not goes back to the ISO boot menu.
The interesting thing is that the BIOS boot sequence has been revised by installer,
The STX installation disk is set as the first boot item.
Best Regards
-------------------
From: Neusoft Team
发件人: Penney, Don <Don.Penney@windriver.com>
发送时间: 2020年5月21日 10:09
收件人: 'starlingx-discuss@lists.starlingx.io' (starlingx-discuss@lists.starlingx.io) <starlingx-discuss@lists.starlingx.io>
主题: [Starlingx-discuss] installer with 4.18 kernel not ejecting ISO on virtual box
Hi folks,
I grabbed the latest ISO built by CENGN and booted it on Virtual Box. It installs fine, but at the end of the installation, it’s failing to eject the ISO (/dev/sr0) and when it reboots, it goes back to the ISO boot menu.
Is anyone else seeing this?
I went back through a number of ISOs, and found that it seems to have started with the update to 4.18 kernel. The April 30th ISO has the old kernel in the installer and is fine:
but when I try the next available ISO, May 5th, I start seeing this problem:
My designer builds are using an older installer, so I hadn’t hit this until trying the CENGN build (which updates the installer with the newly built kernel on each build). The kickstarts would be the same between my designer
build and the latest CENGN build, so I think that just leaves the kernel change in the installer.
At first, I thought it was the following error, seen on the console just before the reboot, but I see it in both the 3.10 and 4.18 cases:
[ 819.009060] systemd-shutdown[9055]: Failed to unmount /run/install/repo: Device or resource busy
I verified the kernel version in the installer by hitting ctrl-o while anaconda is installing, to get to a shell, then running “uname -a”.
Obviously, this would only be an issue for ISO boots, not a network install. Has anyone tried installing the CENGN ISO with the 4.18 kernel on hardware from USB, or have all installs been from network? Maybe this is just
an issue on Virtual Box.
I’m running currently running Virtual Box 6.0.4 r128413 on Windows 10, if that matters. I’ve also tried the latest CENGN ISO on Virtual Box 5.2.22 r126460 on Windows 7 and see the same behaviour.
Don Penney, Developer,
Wind River