NOTES
The notes below have been combined from CentOS 8 Migration Notes and CentOS 7 Migration Notes.
It is based on the Xfce Desktop Environment. MATE is available, see notes at bottom of the page.
Before Starting
/etc/hosts/etc/samba/smb.conf/var/lib/samba/private/passdb.tbd/var/lib/samba/private/secrets.tbd/etc/vsftpd/*Installation
dnf install net-tools and record IP address./etc/selinux/config and change from enforcing to permissive. Reboot.Change the system hostname
nmcli general hostname nmcli general hostname new_server_name systemctl restart systemd-hostnamed
~/.config/xfce4/ from another pre-configured machine to the target machinecd .config/xfce4; rsync -avzn --delete --progress . XX.XX.XX.XX:~/.config/xfce4/ ~/.config/Thunar/ from another pre-configured machine to the target machinecd .config/Thunar; rsync -avzn --delete --progress . XX.XX.XX.XX:~/.config/Thunar/nmcli radio wifi offdnf install iwl6000g2b-firmware.noarchmate-user-admin is a graphical admin tool (old tool was system-config-users)/etc/httpd/conf/httpd.conf (and all backup versions) from previous machine/etc/php.ini file from previous machinefirewall-config to clone rules from previous machinefirewall-cmd --zone=public --add-service=http --permanentfirewall-cmd --zone=public --add-service=https --permanent/etc/fail2ban/jail.local to ban for longer than default of 1 hour/var/log/fail2ban.log outputsystemctl enable fail2bansystemctl restart fail2ban-all-databases option and is an error in link abovepip install dictor and pip install astralsmbpasswd -a USER (then enter password at next prompt)testparm -vsetsebool -P allow_ftpd_full_access onsystemctl restart vsftpdsystemctl enable vsftpddnf -y install ddclient/etc/ddclient.conf for configs from previous machine/var/www/html/[email protected] so that emails sent to localhost are actually delivered to the sys adminroot send email to my actual email address/etc/environment to include LC_TIME="en_GB.UTF-8" for 24-hour clock used by cron jobs$HOME/.local/share/remmina and $HOME/.config/remminasystemctl restart cupsFoomatic/hl1250 enlpstat -p -d and/or set it with lpoptions -d PRINTER_NAMElpstat -d should now show the new printer as the default/usr/bin/sunwait -v sun down -0:01:00 33.640411N 84.419853W/etc/yum.repos.d/* configuration setupdnf install speedtest-clidnf localinstall passwordsafe-fedora37-1.16.rpmcd ~/.local/share/keyrings; mv login.keyring login.keyring.IGNORE Post Installation
/etc/default/grub by setting GRUB_TIMEOUT to the desired value.grub2-mkconfig -o /boot/grub2/grub.cfgalpine to check email from crons that indicate any errors or failuresfastestmirror=1max_parallel_downloads=8Other Notes
MATE is now available. See MATE Documentation
dnf -y install mate-system-monitorKnown issues
2023-11-20
Upgrading Rocky Linux 9.2 to 9.3 in late November 2023 resulted in ImageMagick errors. Tried to remove ImageMagick*, then upgrade 9.2 to 9.3, and then reinstall ImageMagick* but the same error occurs. Fix was to disable /etc/yum.repos.d/epel-next.repo and then reinstall ImageMagick*.
2024-05-13
Upgrade to Rocky Linux 9.4:
Had to run dnf remove gnome-applets to allow upgrade to proceed.
After the upgrade the test Rocky Linux 9.3 VM would not start X with Xfce. Tried a bunch of things like reinstalling all RPMs, getting list of matching RPMs against another VM that upgraded OK but X still would not start. /var/log/messages just kept showing the X server was crashing.
Discovered that if you select the "Standard (Wayland display server)" option on the login page, let the system start its display and log out, then you can login with Xfce successfully.
Very odd but that seems to be the workaround for this upgrade.
2024-11-21
Upgrade to Rocky Linux 9.5:
Had to run the following due to package problems (which will break 'vlc')
dnf remove compat-ffmpeg4 ffmpeg-libs vlc dnf update dnf install vlc gstreamer1-plugin-openh264
Then security cameras running vlc and totem worked correctly.
2024-11-25
Old news, but LibreOffice no longer supported by RedHat. Get the latest LibreOffice from https://www.libreoffice.org/download/download-libreoffice/
Unpack the tar and verify you see individual RPMs. Then:
dnf localinstall *.rpm
You can then start an application like this:
libreoffice24.8 --calc test.csv
Make a .bashrc alias for shortcuts.
2025-06-25
Upgrade to Rocky Linux 9.6 initially failed on test VM with package errors.
Trying these steps on a test VM:
dnf remove vlc totem gstreamer1-plugins-bad-free gstreamer1-plugins-ugly-free dnf update dnf install vlc totem
Then reinstall VBox Guest Additions.
Got message /var/log/vboxadd-setup.log to find out what went wrong (Building the modules for [current] kernel failed).
Rebooted and then tried installing kernel-debug-devel but same problem.
Upgraded host computer's VirtualBox to version 7.1-7.1.10 (was running 7.1.4)
Now can upgrade from 9.5 to 9.6 on VirtualBox VMs but Xfce logs in for 1 second, a black screens appears, and then the sytem logs out.
The fix is to choose a different window manager on the main login screen (upper right, choose "GNOME Classic on Wayland"). After logging in successfully, log out and then go back to the main screen and choose the "Xfce Session".
For my non-virtual, primary Rocky Linux 9 machine the upgrade from 9.5 to 9.6 worked without issue but upon reboot, I got the following error:
VirtualBox can't operate in VMX root mode. Please disable the KVM kernel extension, recompile your kernel and reboot (VERR_VMX_IN_VMX_ROOT_MODE).
Using a Google Search, the temporary fix was to do the following see PDF:
# lsmod | grep kvm
Verified the following output:
kvm_intel 446464 0 kvm 1404928 1 kvm_intel
Disabled the KVM temporarily:
# modprobe -r kvm_intel
Verified nothing was returned with the following command:
# lsmod | grep kvm
Was then able to start up the VMs using the VirtualBox GUI successfully.
Need to reboot machine and see if fix stays. Otherwise, the permanent fix is the following:
Create a file named kvm-blacklist.conf in /etc/modprobe.d/
Add the blacklist entry: Add the appropriate line to the file:
blacklist kvm_intel
-or-
blacklist kvm_amd
Reboot to apply the changes.
2025-10-18
The network manager icon in the system try disappeared. Had to do the following:
dnf install network-manager-applet
Go to Applications > Settings > Session and Startup
Then Application Autostart > Add
Command: /usr/bin/nm-applet
Trigger: on login
Reboot and the icon stayed.
2025-10-20
Some kernels were missing from DNF repos; Virtual Box would not start after upgrading. Downgraded to previous kernel, rebooted twice, and then it started OK.
2026-12-05
For Rocky Linux 9.7, trying to login to Xfce after initial install may not work (immediate logout). If that happens, login to a standard Wayland display session. Then log out and login to Xfce again.
2026-05-29
Upgrade to Rocky Linux 9.8:
Initial reboot on the 'rl9-test-sel' VM did not work. Needed to power off VM and reboot again.
Other machine produced this error:
# reset_aide
ERROR: /etc/aide.conf:7: unexpected character: ':' (line: 'database=file:@@{DBDIR}/aide.db.gz')
mv: cannot stat '/var/lib/aide/aide.db.new.gz': No such file or directory
/usr/sbin/aide: unrecognized option '--verbose=2'
Upon investigation, Aide was upgraded from version 0.16 to version 0.19.2
The new command should be one of the following:
/usr/sbin/aide --check --log-level=info /usr/sbin/aide --check --log-level=debug
Also the baseline /etc/aide.conf was updaded. Compare my changes I made from the original 0.16 config and merge them into the new 0.19.2 config.
Major problem on my host machine - VirtualBox will not start any VMs. The error is:
WARNING: The vboxdrv kernel module is not loaded. Either there is no module
available for the current kernel (5.14.0-687.10.1.el9_8.0.1.x86_64) or it failed to
load. Please recompile the kernel module and install it by
sudo /sbin/vboxconfig
You will not be able to start VMs until this problem is fixed.
VBoxHeadless: Error -1908 in suplibOsInit!
VBoxHeadless: Kernel driver not installed
VBoxHeadless: Tip! Make sure the kernel module is loaded. It may also help to reinstall VirtualBox.
See https://github.com/VirtualBox/virtualbox/issues/676 for a discussion and apparent fix in the VirtualBox test builds. VirtualBox-7.2-7.2.9_174006_el9-1.x86_64.rpm installs and runs all VMs correctly.
Later on had a problem with xrdp-selinux (simply removed the package as I was not using it)
Also had to migrate perl-DBD-MySQL to perl-DBD-MariaDB. Did the following:
dnf remove perl-DBD-MySQL
That also removed mytop, mysql-common and mysql-libs. Updated a few .pl scripts to replace $driver = "mysql" with $driver = "MariaDB".
For mytop you can use the built-in client utility mysqladmin. For an exact mytop clone experience, you can wrap it in a terminal watch loop by executing watch -n1 "mysqladmin processlist -u root -p".
2026-06-08
When using Chrome, the CPU fan is ramping up (annoying). With 3 VMs running it seems the machine is just getting to the point of having a lot of processing going on all the time.
ChatGPT says to "Limit turbo boost. The fan ramping is frequently caused by the CPU briefly jumping from 1.9 GHz to 4.6 GHz. You can inspect current settings:"
cat /sys/devices/system/cpu/intel_pstate/no_turbo
Temporarily disable turbo:
echo 1 | sudo tee /sys/devices/system/cpu/intel_pstate/no_turbo
Re-enable:
echo 0 | sudo tee /sys/devices/system/cpu/intel_pstate/no_turbo
"Many users find that disabling turbo dramatically reduces fan fluctuations while having little impact on ordinary web browsing."
Made the change on the evening of 8 June 2026 and it immediately made the fan speed stop jumping up and down. A few days aga I also set the VirtualBox VMs to consume no more than 50% CPU but that didn't seem to make any difference.
As a last resort, try the limiting the CPU consumed by chrome with the following:
# Install cpulimit sudo dnf install cpulimit -y
# Run Chrome with a CPU cap (e.g., 50% across all cores) cpulimit -e chrome -l 200 &
Just kill the cpulimit process when you want to stop throttling Chrome.