• If you are still using CentOS 7.9, it's time to convert to Alma 8 with the free centos2alma tool by Plesk or Plesk Migrator. Please let us know your experiences or concerns in this thread:
    CentOS2Alma discussion

Question IPs disconnected after plesk updates

andyxyz

Basic Pleskian
HI there

About once a month across multiple servers all IPs other than the default are disconnected.

The only way to fix is to log into plesk. > tools > IP address > rescan > repair all ip then run plesk repair web -y- v

This morning 20 servers were affected taking all sites associated with those server off line.

This has been happening for 6 months or so. It is not always the same servers that are affected.

All servers appear to be running Ubuntu 18.04

Anyone got any ideas?
 
After yesterdays updates configuration of all three IPv6 is broken too on Ubuntu 20.04.2 LTS and Obsidian latest relaese.
etc/network/interfaces was changed automatically by Plesk/System.
What is going on? I changed configuration back to teh old, but websites can't connect by IPv6, only by IPv4.
But I use IPv6 for own domain related nameservers. Help would be great.
 
~
etc/network/interfaces was changed automatically by Plesk/System.
What is going on? I changed configuration back to teh old, but websites can't connect by IPv6, only by IPv4.
~
You're using IONOS, Plesk & Ubuntu 20.02.2 LTS just like us (but we're using Cloud Servers not Dedicated Servers like yourself). We've never, ever seen Plesk change the etc/network/interfaces file in all the time we've used IONOS, Plesk & Ubuntu but, we did have an early issue when using IPv6 at IONOS with Ubuntu 20.04. After some 'trial and error' testing, we determined this was related to using IPv6 addresses within the /etc/systemd/resolved.conf & the /run/systemd/resolve/resolv.conf file configs. We changed those & we haven't had the IONOS & Ubuntu 20.02.2 LTS IPv6 issue since, for what it's worth.
 
OK so we had another server die due this again today

IT is very odd. I know at one point plesk were adding new IPs through netplan. However, this seesm to have revered back to /etc/network/interfaces

and it seems to replicate the ips over and over



cat /etc/network/interfaces
# interfaces(5) file used by ifup(8) and ifdown(8)
# Include files from /etc/network/interfaces.d:
source-directory /etc/network/interfaces.d
iface eth0:1 inet static
address xxx.xxx.xxx.142
netmask 255.255.255.255
auto eth0:1
iface eth0:2 inet static
address xxx.xxx.xxx.143
netmask 255.255.255.255
auto eth0:2
iface eth0:3 inet static
address xxx.xxx.xxx.178
netmask 255.255.255.255
auto eth0:3
iface eth0:4 inet static
address xxx.xxx.xxx.215
netmask 255.255.255.255
auto eth0:4
iface eth0:1 inet static
address xxx.xxx.xxx.142
netmask 255.255.255.255
auto eth0:1
iface eth0:2 inet static
address xxx.xxx.xxx.143
netmask 255.255.255.255
auto eth0:2
iface eth0:3 inet static
address xxx.xxx.xxx.215
netmask 255.255.255.255
auto eth0:3
iface eth0:4 inet static
address xxx.xxx.xxx.178
netmask 255.255.255.255
auto eth0:4
iface eth0:1 inet static
address xxx.xxx.xxx.142
netmask 255.255.255.255
auto eth0:1
iface eth0:2 inet static
address xxx.xxx.xxx.143
netmask 255.255.255.255
auto eth0:2
iface eth0:3 inet static
address xxx.xxx.xxx.178
netmask 255.255.255.255
auto eth0:3
iface eth0:4 inet static
address xxx.xxx.xxx.215
netmask 255.255.255.255
auto eth0:4
iface eth0:1 inet static
address xxx.xxx.xxx.142
netmask 255.255.255.255
auto eth0:1
iface eth0:2 inet static
address xxx.xxx.xxx.143
netmask 255.255.255.255
auto eth0:2
iface eth0:3 inet static
address xxx.xxx.xxx.178
netmask 255.255.255.255
auto eth0:3
iface eth0:4 inet static
address xxx.xxx.xxx.215
netmask 255.255.255.255
auto eth0:4
iface eth0:1 inet static
address xxx.xxx.xxx.142
netmask 255.255.255.255
auto eth0:1
iface eth0:2 inet static
address xxx.xxx.xxx.143
netmask 255.255.255.255
auto eth0:2
iface eth0:3 inet static
address xxx.xxx.xxx.178
netmask 255.255.255.255
auto eth0:3
iface eth0:4 inet static
address xxx.xxx.xxx.215
netmask 255.255.255.255
auto eth0:4
iface eth0:1 inet static
address xxx.xxx.xxx.142
netmask 255.255.255.255
auto eth0:1
iface eth0:2 inet static
address xxx.xxx.xxx.143
netmask 255.255.255.255
auto eth0:2
 
iface eth0:1 inet static
Strange, using eth0:x to assign additional addresses is the old way to do so (should be "ip addr add" now). Also it is suspicious that there is no assignment for eth0 itself. Is it possible that this configuration was originally from a much older system?
 
Standard install of ubutnu 18.04 (ip assigned through netplan at install as usual) then install of plesk
all new IPs (added in plesk) are created using the old way.. then get disconnected on upgrades
 
Hi @ all friends of Plesk,
Hi @learning_curve, to the Plesk support team I reported (and discussed) the problem "Adding multiple IP configurations (IPv4 + IPv6) in /etc/network/interfaces by Plesk system on Ubuntu 20.04 server when you click on Repair IP in Plesk IP settings".
One day I experienced having the same configuration for several IPs multiple times (one IP had 4 times the same configuration, better to describe as cloned configurations).
For this issue I made a support ticket, gave access to my server, and it was confirmed as bug PPPM-12941 by Plesk support on 22th of April in 2021.
It has to do with the IP repairing system. But they also discovered that there is another problem with misconfiguration in etc/network/interfaces for which Ionos is responsible. Two bugs brought my server offline every 6 to 12 hours.
As for changing the network configuration so only one of either Netplan or ifupdown is used - the tool for managing the network is chosen on the system level, and it is done on all Linux servers, regardless of whether Plesk is installed. When Plesk is installed on a server, for adding or deleting IP addresses to the network interfaces, it just uses the tool that is already used in the system, it does not manage the lower-level configuration of which tool is used for managing the network in the system. In the moment both are active/running: Netplan "and" ifupdown, this seems to be the big problem.

THE PLESK-IONOS-UBUNTU-SOLUTION:
I was able to solve all problems together by only returning to the default network/interfaces configuration with the main IPv4 from provider Ionos. Removing all additional IPs (two IPv4 and three IPv6), only the main server IPv4 existing. Much work to edit all subscriptions and DNS (I use the additional IPs for own domain nameservers).
In former times it was absolutely necessary to edit network/interfaces for using additional IPs, but not in present times. Today you only have to add the IPs over Plesk interface, NO editing network/interfaces. Then re-do all the changes in subscriptions and the IPs with their corresponding nameservers. So, in the end, server is running fine and smooth again with no need to repair IPs.
Greets
 
Last edited:
@Dukemaster You've had issues with IONOS / Ubuntu / Plesk which we haven't. Most likley, because you're using multiple IP addrresses per server, but we're not. We only use one IPv4 and one IPv6 address per IONOS Cloud Server. Consequently, we've never had to use Repair IP in Plesk IP settings as you have.

We discovered the IONOS use of "ifupdown" and NOT "Netplan" to fit in with their own Network Interfaces / DNS / Nameserver preferences etc in their Cloud Servers combined disk image quite a long time ago, when we first switched to Ubuntu 18.04 and have posted about it, a few times since, in this very forum.

IF you know about this "config change" in advance and IF you have modified these two files: /etc/network/interfaces and /etc/sysctl.conf so that IPv6 is fully functional with with the IONOS / Ubuntu / Plesk Cloud Servers combined disk image, everything works, exactly as expected and just as you have mentioned (i.e. Plesk using the existing system tool etc) IF you don't, there will be some problems with IPv6 when using that IONOS Cloud Servers combined disk image.

The other issue that we mentioned in post #3 above, was also related to IONOS & IPv6 but it is totally different than the one that you described in your initial post. It was added as a "just in case", you being a fellow IONOS user and all :)
 
Yo, @learning_curve, it was not good to come in this thread with another problem. But I must say, when I started I thought that one of my problems is deeply related to the problem of the thread starter. I apologize. I was not the only one who mixed the topics. This is not good, but so typical for forums. :D
 
We lost another 25 servers to an update this morning. Has anyone got any idea how we can fix this. It seems really odd that others are not experiencing the same issue on what must be a very common combo.. ie ubuntu 18.04 / plesk
 
I have the same problem, today round 6:26 cest our server gone down, had to click repair ip adresse.... i only use ipv4 so i disabled ipv6 now. but i dont know if this problem will be gone by that.... i noticed that at the same time there was some system updates by plesk. ionos rootserver, ubuntu 20.04 lts
 
We're running Ubuntu 20.04.2 LTS everywhere and have upgraded to Obsidian 18.0.37 without any issues at all.
Maybe we're lucky, but we haven't had any big issues (after Plesk upgrades), since we started using Plesk.
So, No complaints from us!
 
From what it looks like, this problem is isolated to Ionos servers? We run CentOS so I wouldn't know lol.
 
No these are standalone dedicated servers.. nothing to do with ionos

Does NOT affect VMware servers at all
 
Happened again, again 6:25 cest, all ips other that the main ones are disconnected, all sites down :-(
 
Back
Top