Cloud-init and ifcfg-eth0

[Alexander86]

Hi, I have a template on which cloud init is installed, I do fullclone from it and after that I change the ip and gw parameters.
Option 1 - qm set 123 --ipconfig0 ip = 10.0.10.123 / 24, gw = 10.0.10.1
Option 2 Post request / nodes /% s /% s /% d / config in the body of which new parameters ip, gw are transmitted.
In ui Proxmox, the new parameters change, in the virtual machine configuration in the vmid.conf file, the parameters change to new , in the qocow2 file are changed to new parameters, but in the file etc / sysconfig / network-scripts / ifcfg-eth0 the parameters are not written. Tell me how to solve the problem thanks?

 

[mira]

Can you provide your VM config? (The ouput of qm config 123)

 

[Alexander86]

Can you provide your VM config? (The ouput of qm config 123)

boot: dcn
bootdisk: scsi0
citype: configdrive2
cores: 2
ide2: local:105/vm-105-cloudinit.qcow2,media=cdrom,size=5M
ipconfig0: ip=192.168.236.109/22,gw=192.168.238.2
localtime: 1
memory: 6144
name: 23674--current
net0: virtio=E6:E7:A9:85:6B:7E,bridge=vmbr0
numa: 1
ostype: l26
scsi0: vm-zvol-4:vm-105-disk-1,discard=on,iothread=1,size=60G
scsihw: virtio-scsi-single
smbios1: uuid=3e587edc-2c0f-487c-b669-27ebe9928c14
sockets: 2
vga: qxl

 

[mira]

You should not specify configdrive2 for a linux VM, see https://pve.proxmox.com/pve-docs/pve-admin-guide.html#qm_cloud_init
You can change that via qm set 123 --citype nocloud

 

[Alexander86]

You should not specify configdrive2 for a linux VM, see
You can change that via qm set 123 --citype nocloud

i know, but this was done after the problem was discovered (( just for the sake of interest %)

 

[Alexander86]

this is template params :
boot: dcn
bootdisk: scsi0
cores: 2
ide2: local:23674/vm-23674-cloudinit.qcow2,media=cdrom
ipconfig0: ip=192.168.236.74/22,gw=192.168.238.1
localtime: 1
memory: 6144
name: Sniffer-Alexander
net0: virtio=F2:A8:44:E3:7D:76,bridge=vmbr0
numa: 1
ostype: l26
parent: clean-7-0-10
protection: 1
scsi0: vm-zvol-4:base-23674-disk-1,discard=on,iothread=1,size=60G
scsihw: virtio-scsi-single
smbios1: uuid=95568f8d-1e27-424f-b644-da8f5e416255
sockets: 2
template: 1
vga: qxl

 

[mira]

Which distribution are you using? And which version of cloud-init do you use?

 

[Alexander86]

Which distribution are you using? And which version of cloud-init do you use?

HI, use centos 7 and cloud-init 0.7.9.

I opened the logs and the stages.py error appeared in them, in more detail in the screenshot. When I start the machine after executing the
"pvesh set /nodes/237-197/qemu/110/config --ipconfig1 ip=192.168.236.109/22,gw=192.168.236.132"
request, next to the main settings of the network ens18, a file with eth0 starts appears, if I change the name of the file ens18 to eth0 (as well as the settings inside the file to eth0), then the system does not update the information in it.

 

[Alexander86]

Screenshots from net-settings

 

[mira]

Your cloud-init version might be too old, can you try using a newer package? (17.x or if possible a 18.x version)

 

[Alexander86]

Your cloud-init version might be too old, can you try using a newer package? (17.x or if possible a 18.x version)

View attachment 8626

When cloud-init makes changes in the network configuration file, it is written in IPPADDR = 0.0.0.0 /00 format, but it must write the ip address and prefix separately. Is this kind of problem you encountered? Naturally, the network does not work when everything in the file is written in one line.

 

[mira]

Your cloud-init version is simply too old, it doesn't conform to the network configuration defined by cloud-init. Either update cloud-init, choose a different distribution with a newer cloud-init version (Fedora 28 for example, or 29) or choose a different method to configure your network.

 

[Alexander86]

Your cloud-init version is simply too old, it doesn't conform to the network configuration defined by cloud-init. Either update cloud-init, choose a different distribution with a newer cloud-init version (Fedora 28 for example, or 29) or choose a different method to configure your network.

I give up, the fact is that I can use centos 7 for work, and in yum it puts only 0.7.9 version, how to build and install a package with version 18 under centos 7 I don’t know yet if you have any hint I will be very grateful to you

 

[mira]

You could try installing the Fedora 28 or 29 package. (Should be able to download it from here: https://apps.fedoraproject.org/packages/cloud-init ) Still only 17.1 but network configuration worked with Fedora 28

 

[Alexander86]

You could try installing the Fedora 28 or 29 package. (Should be able to download it from here: https://apps.fedoraproject.org/packages/cloud-init ) Still only 17.1 but network configuration worked with Fedora 28

Thanks I try =)

 

[Alexander86]

Hello again, when installing cloud-init, a dependency error occurs. The machine installed Python 2.7 as well as versions 3.4 and 3.6, all of these versions have been downloaded via pip install necessary dependencies. They are located in the /usr/lib/python ... folder. But the error still repeats.


[root@localhost site-packages]# pip3 install six
Collecting six
Using cached https://files.pythonhosted.org/pack...863fe3056e86a/six-1.11.0-py2.py3-none-any.whl
Installing collected packages: six
Successfully installed six-1.11.0

[root@localhost site-packages]# rpm -i /Developing/cloud-init-17.1-7.fc29.noarch.rpm
error: Failed dependencies:
libselinux-python3 is needed by cloud-init-17.1-7.fc29.noarch
policycoreutils-python3 is needed by cloud-init-17.1-7.fc29.noarch
python(abi) = 3.7 is needed by cloud-init-17.1-7.fc29.noarch
python3-PyYAML is needed by cloud-init-17.1-7.fc29.noarch
/..................................................................................................../
python3-six is needed by cloud-init-17.1-7.fc29.noarch

[root@localhost site-packages]#