View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0017877||CentOS-8||dnf||public||2020-11-22 04:39||2020-11-25 17:21|
|Platform||Bare metal x86-64||OS||CentOS||OS Version||8|
|Target Version||Fixed in Version|
|Summary||0017877: Receive error message from/about DNF in POSTTRANS script during installation of kernel-core RPM; INSTALLATION STOPS.|
|Description||First of all, isn't "installation" a reasonable "Category" to have for bug tracking? If it's there I can't see it. I notice the other reports related to this problem used "kernel" and "anaconda". I don't think those are accurate descriptions of the nature of the problem. I don't understand why there aren't more reports of this symptom and why the other reports of this symptom aren't treated more seriously. The suggestions to add meaningless parameters to the end of a virtualized install command line don't help much if you are not using a command line nor installing to a virtualized environment. Okay, I'll calm down. After losing my previous CentOS 7.something installation to the after effects of the "boothole" fix, I thought I would simply start over with the latest CentOS 8. Downloaded the file, wrote it to a USB, booted it directly on the bare metal of my laptop and went through the straightforward choice of a graphical installation. Chose Workstation plus some other features. Actual installation start and runs. Consistently I will eventually receive a pop-up window announcing that DNF has encountered a problem on the POSTTRANS script while installing the kernel-core RPM. The only obvious option is to click a button marked EXIT. The installation is incomplete to some unknown extent. Rebooting the install media into repair mode has allowed me to experiment. Explicitly installing kernel-core from the install media worked but did not result in a bootable system. Explicitly upgrading kernel-core from online repositories succeeded but did not produce a bootable system. "dnf distro-sync" DID produce the bootable system I am suffering with now and using to enter this angry report. I say "suffering" because the performance is really slow/unresponsive in starting any graphical application AND the Terminal application never starts and shows me a window. I click on "Activities" and a vertical menu of likely applications appears. I click on the "Terminal" application icon and a tab appears next to the "Activities" tab and a white half-circle chases its tail for a count of one-thousand-thirteen and the tab goes away and no "Terminal" window appears. Files, Firefox, and Chrome run as expected but startup always takes a I have been attributing this behavior to actions that should have happened after the installation of "kernel-core" never happening because the installation process never gets that far. I see other reports that command-line driven (text-mode) installations don't have this POSTTRANS problem. How can that be? I may give it a try and report here.|
|Steps To Reproduce||Obtain CentOS 8.2 2004 installation material.|
Write to USB,
Boot from USB.
Choose Installation from boot menu.
Make reasonable choices from the installation options including installation to real (unfortunately spinning rust) hard drive.
Wait patiently as lots of processing takes place.
Gasp in alarm as message about DNF encountering an error during POSTTRANS script of kernel-core rpm.
Reluctantly click "Exit" button.
Try to salvage the effort just invested.
Give up after a week and enter problem report into CentOS Bug Tracker.
|Additional Information||After system has been wrestled into some semblance of functionality:|
One graphical application (Terminal) simply will not start
Other graphical applications, Files, Firefox, Chrome consistently take 10+ seconds to start every time.
I just can't believe that the type of installation I'm trying to do is anything out side the mainstream. Why isn't everybody having this problem and reporting it?