Applicable to:
- Plesk for Linux
Symptoms
- Debian 11 server (most often Debian 11.10) is down and cannot boot after Plesk auto-update
- If the server in question becomes accessible to a point at which it can boot, it becomes evident that Plesk core packages are gone:
# root@server:~# apt list --installed | grep plesk-core
Simple commands that depend on
root@server:~#systemdfunctioning properly are also not working, which suggests that other core OS packages are also missing:# root@server:~# hostnamectl
System has not been booted with systemd as init system (PID 1). Can't operate.
Failed to connect to bus: Host is down - The Imunify360 extension (and its repositories) is installed on the server
Cause
The issue is tied to a known bug with ID #PPPM-14558. The bug will be resolved in future versions of Plesk Obsidian.
You may track when this bug will be resolved in the Change Log for Plesk Obsidian.
Under specific circumstances, the PUM (Plesk Update Manager) tool refuses to update libkrb5* packages, because it was installed from the oldstable repository, but the candidates are from the oldstable-security repository, which is not compatible with the oldstable repository. At the same time, some packages (mostly Imunify360 packages) are selected for upgrade, with hold of libkrb5* packages ongoing, which results in the improper removal of many packages from the server.
Resolution
Warning: If you control Debian 11 servers that have not yet been affected by this issue, you should make sure that you update all of their packages to their latest versions by running the apt update && apt upgrade command instead of using the pum Plesk tool.
Since the Debian OS core is affected severely from the issue, if you have a server snapshot available, it is suggested that you restore the server from that snapshot.
If a server snapshot is not available, you need to execute the disaster recovery process to restore the website data and database data from the old server to the new server by using the information in this article:
How to restore a Plesk server on a new Linux server from file system?
After this process has been finalized successfully, you should restore the Plesk database from the last pre-upgrade dump that is present on the old server's file system by using the steps in this article:
How to restore Plesk system databases from a Plesk daily/pre-upgrade dump
Comments
I encountered this error when upgrading from Debian 11 to 12 (Plesk 18.0.75).
Systemd simply won't start anymore.
The solution is complex:
Start the server with a different Linux distribution or rescue system.
Then bind the system partitions:
mkdir /system
mount /dev/sda3 /system #(sda3 is SystemPart for me)
mount -o bind /dev /system/dev
mount -o bind /proc /system/proc
mount -o bind /sys /system/sys
chroot /system
mount /dev/sda2 /boot #(boot partition)
mount /dev/sda5 /var #(I had separate partitions for /var.)
edit /etc/default/grub;
GRUB_CMDLINE_LINUX_DEFAULT="quiet init=/lib/systemd/systemd" #(Plus your parameters)
grub-install /dev/sda
update-grub2
If there is a problem, remove "init=/lib/systemd/systemd" and repeat the procedure.
Please sign in to leave a comment.