Issue Autoupdate Plesk 18.0.56 failed causing major problems

[clemensmol]

Server operating system version

AlmaLinux 9.3

Plesk version and microupdate number

18.0.56

Server is completely down, no network connection possible anymore. DBus service wouldn't start anymore, crashing on some empty config files. I removed those, dbus is running again. But I can't get the server online anymore, "network unavailable" errors all over the place.

 

[Peter Debik]

The Plesk update does not change or update system services. Did you check the /var/log/plesk/install/autoinstaller3.log and what errors do you see there that lead you to the conclusion that Plesk caused issues on the operating system? Maybe the issues of the operating system caused the Plesk update to fail?

 

[clemensmol]

I came to this conclusion by reading the autoinstaller3.log and /var/log/messages. Of course I don't know for sure its Plesk, but the problems started duren/after an automatic Plesk update. Could also be a faulty plugin?

Plesk autoinstaller was downloading files succesfully and started updating components. Last "succesfull?" lines are at 04:17, than a blank line and around 04:22 the autoinstaller continues? but then all of the sudden it shows errors "could not resolve host".

Between 04:17 and 04:22 the server seems to be rebooting and I find errors in /var/log/messages

dbus-broker-launch[9411]: INVALID XML in /etc/dbus-1/system-d/org.fedoraproject.SetroubleshootPrivileged.conf +1: noe element found

failed to start D-Bus System Message Bus

A few lines below errors like this one pop up:

network unreachable resolving './DNSKEY/IN':

 

[Peter Debik]

The first line quoted ("SettroubleshootPrivileged.conf") points to an issue with SELinux. The last line is probably just a symptom. No D-Bus = system won't work properly, so all the following errors are probably caused or linked with the first.

I can say that I updated Alma from 9.2 to 9.3 on a production machine this morning, with SELinux and SELinux Plesk context active, and it did not cause any issues. It had updated 207 files of the operating system.

If you suspect that Plesk, the update respectively, is the cause for the D-Bus error, you could submit a ticket to Plesk support and let a support engineer check it. From my personal experience such issues are rather results of disks delivering wrong or uninterpretable data (maybe there is an error on the disk) or some awkward, rare other glitches on the OS level. Maybe an older Kernel running along with newer other services.

 

[clemensmol]

Thanks for replying Peter. I hired an expert to fix the problems, we're not quit sure but we think there were "changes" somewhere in the system which were applied after rebooting. The plesk autoupdater probably triggered the reboot. It might be related to an AlmaLinux update (/etc/sysconfig/network-scripts isn't used anymore).

 

[Maarten.]

The Plesk autoupdater never issues or triggers a reboot. Can you examine the /var/log/messages file and the date/time just before the reboot? There must be something that tells you what caused the reboot of the server.

 

[jwvdmei]

Hello, I think I have a similar problem.

The server kept crashing and CPU's maxed out. First I added more RAM but that didn't solve it. Then I thought maybe one of the following: DDOS, a runaway script, bad WP plugin, hacked etc. Eventually I got so sick of it all, I put back a snapshot from two days previous when everything was working fine, assuming that should solve the problem, however the problem persisted. After about the 100th time rebooting and not seeing strange processes with "# top" it occurred to me that maybe Plesk is trying to update (automatically) but got stuck in doing so. I have now turned the automatic update feature off and the system seems to be working nicely again. Not a solution but some temporary relief so one can enjoy their free time again.