Re: systemd - nspawn ; systemd-networkd ; containers

 I get the error when, on the host, I am creating the container.  

 

  To this point, I have found my setups work needing systemd-networkd on both the host and container.

  

  The distro for the container I am working to set up is 'Fedora 30 '.  That is also

the host distro too.  Seems I have something wrong on my side. 

 Thanks for your help/time.  What else can I provide that will aid in isolating this?

 

 I have running CentOS 7 and Fedora 29 hosts that have a mix of fedora and centos containers

running with their own unique IPs and the goal ==> migrate to Fedora30 and CentOS8

//From notes I have captured for my setup:

//must run systemd-networkd

 systemd-networkd is a system daemon that manages network configurations.

  It detects and configures network devices as they appear; it can also

  create virtual network devices. This service can be especially useful

  to set up complex network configurations for a container managed by

  systemd-nspawn or for virtual machines. It also works fine on simple connections.

On Fri, Sep 27, 2019 at 10:31 PM Ryan Gonzalez <rymg19@xxxxxxxxx> wrote:

I'm a bit confused, where is the error about networkd not being found coming from? Do you want networkd on the host system or inside the container? If the latter, what distro does the container run?

On Fri, Sep 27, 2019, 6:57 PM Douglas E. Hopley Jr. <hopley@xxxxxxxxxxxxx> wrote:

Greetings -  I hope this finds you well and that I provide a set of good starting details ...

I have been using systemd-nspawn to create 'containers'/machines that I use on hosts.  These machines were set up with unique IP addresses so I have multiple machines on same host (in some cases sharing the host network and in some cases using a secondary ethernet). ...

Recently I updated one of my hosts to F30 and went to create a new machine.

 The main thrust of this email is that in my process of creating a new systemd-nspawn machine it seems that `systemd-networkd` is not installed when I use my steps (I can outline if that will help) to create a machine.  So, in turn I am  not able to set up the machine with unique network configuration at this point.

 I have running systems with systemd v219 (potentially too old, sorry) and v229

that have machines (systemd-nspawn containers) networked.

 At this point I was looking to bring up a new sytemd-nspawn container/machine on a host that has v241. specifically :

==> 

systemd 241 (v241-12.git1e19bcd.fc30)
+PAM +AUDIT +SELINUX +IMA -APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ +LZ4 +SECCOMP +BLKID +ELFUTILS +KMOD +IDN2 -IDN +PCRE2 default-hierarchy=hybrid

 When I follow (my) steps/process for creating a new systemd-nspawn,  I get this response ==> 'No match for argument: systemd-networkd'.

So,  I have  few starting questions and I can provide more details  as needed/requested

1)  Has there been, since v219 or v229 a change (that I have yet to uncover the details) that have moved from systemd-networkd to something new/different/better?  What is that Or what should I be installing instead of systemd-networkd?

2) Can you direct me to the details about how to migrate from 'systemd-networkd'? 

 3) This is a bit of an open point --- Should I use systemd-nspawn machines to run resources?  Is there another suggested path I should educate myself on for having containers (podman? )? I built up my process for bringing up machines and found that very useful/powerful to have  containers running on a host that I can have as unique machines (IP addresses, and processes etc.).  Have I over-used the technology OR is it safe/suggested to use nspawn in this manner?

Later I will look to create more machines on Centos8 (systemctl --version ==>239).

  Let me know if I can add more details Or  if I need to be more specific.

Thanks in advance for your time.