Hello everyone,
I’ve been trying to upgrade my DirectAdmin production server from AlmaLinux 9.6 (Sage Margay) to AlmaLinux 10 using the ELevate (Leapp) migration tool.
Unfortunately, the process has turned into a very complicated situation — several steps failed, and I finally decided to revert the system back to my AlmaLinux 9.6 snapshot.
Here’s the full story for context, hoping someone can provide guidance or confirm if upgrading DirectAdmin-based environments is even safe yet.
System Setup:
-
Control panel: DirectAdmin
-
Web server: LiteSpeed (licensed)
-
Security: Imunify360 + KernelCare
-
OS before upgrade: AlmaLinux 9.6 (Sage Margay)
-
Kernel: 5.14.0-570.46.1.el9_6.x86_64
-
Goal: In-place upgrade to AlmaLinux 10 using ELevate/Leapp
What I Did:
-
Installed the ELevate tools:
dnf install leapp-upgrade leapp-data-almalinux leapp preupgrade -
The preupgrade report listed several issues:
-
“Upgrade is unsupported”
-
“A YUM/DNF repository defined multiple times”
-
“Packages not signed by Red Hat found”
-
“Custom leapp actors detected”
-
“libdb (Berkeley DB) detected on the system”
-
-
I tried to continue by skipping inhibitors:
LEAPP_UNSUPPORTED=1 LEAPP_UNSUPPORTED_INHIBITOR=1 LEAPP_DEVEL_SKIP_RHSM=1 \ leapp upgrade --enablerepo almalinux10-baseos --enablerepo almalinux10-appstream -
During this process, several errors occurred:
-
DNS resolution issues (
curl error 6– could not resolve repo.almalinux.org) -
RPM database corruption (had to rebuild manually):
systemctl stop packagekit rm -f /var/lib/rpm/__db* rpm --rebuilddb -v -
Repositories for AlmaLinux 10 failed to sync (“Cannot download repodata/repomd.xml”).
-
-
After multiple attempts, Leapp gave a “FATAL Actor: rpm_scanner” error and stopped.
-
Even after forcing the upgrade again (
leapp upgrade --reboot), noleapp-upgrade initramfsentry appeared in GRUB.
→ The system rebooted normally into AlmaLinux 9 again.
Result:
After these repeated attempts:
-
Some packages (rpm, python3, systemd) partially updated to EL10 versions.
-
Others stayed EL9.
-
DNS and dnf cache were broken several times.
-
Eventually, the system became unstable — some repos pointed to EL10, some to EL9.
At this point, I restored my snapshot backup of AlmaLinux 9.6, which worked perfectly and brought everything back to normal.
Current State (after snapshot restore):
-
AlmaLinux 9.6, clean kernel
5.14.0-570.62.1.el9_6.x86_64 -
DNF repositories verified — only EL9 repos active:
almalinux-baseos almalinux-appstream almalinux-extras epel crb imunify360 kernelcare -
DirectAdmin and LiteSpeed running fine:
systemctl status directadmin → active (running) systemctl status litespeed → active (running) -
RPM database and all system services are healthy again.
My Questions:
-
Is it currently safe or supported to upgrade DirectAdmin environments (with Imunify360, LiteSpeed, KernelCare) from AlmaLinux 9 → 10?
-
Does the ELevate tool fully support systems running custom or third-party repos like these?
-
Should I expect official support for ELevate → AlmaLinux 10 soon, or is a clean install still the only reliable path?
-
Is there a tested or community-proven method to perform this migration on production hosting servers?
Summary:
-
Tried AlmaLinux 9.6 → 10 upgrade via ELevate
-
Encountered repo duplication, RPM corruption, DNS errors
-
Leapp process never reached the initramfs stage
-
Restored to clean 9.6 snapshot (system is now stable again)
-
Seeking advice on the proper and safe method for DirectAdmin servers
Thanks in advance for any insight!
If logs are needed, I can provide /var/log/leapp/leapp-report.txt and /var/log/leapp/leapp-upgrade.log.
– Can