Uploaded image for project: 'RHEL'
  1. RHEL
  2. RHEL-67815

Kickstart deployment doesn't start if second nic is connected but normally unconfigured

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Blocker Blocker
    • None
    • CentOS Stream 10
    • anaconda
    • None
    • None
    • None
    • rhel-sst-installer
    • ssg_front_door
    • None
    • None
    • None
    • None
    • None
    • x86_64
    • None

      What were you trying to do that didn't work?

      Deploying CentOS Stream 10 for https://testing.stream.centos.org setup.

      We use ansible automation and for bare-metal deployments, we automatically reinstall supermicro blades with two nics. that's the same infra that we used to validate our centos  linux 7 / stream 8 and stream 9 composes but it fails on centos stream 10 (more info below)

      What is the impact of this issue to you?

      not able to validate centos stream 10 (upstream of RHEL) composes

      Please provide the package NVR for which the bug is seen:

      unable to say if that's really anaconda or else so feel free to reassign but current centos stream 10 composes aren't installable due to the bug :

      How reproducible is this bug?:

      always, as long as a physical machine has two nics reporting link and even if kickstart should only configure first interface.

      From pxe config :

      LABEL cody-n27.ci.centos.org-deploy
              MENU LABEL Deploying cody-n27.ci.centos.org - kickstart (OS 10-stream/x86_64)
              kernel boot/centos/10-stream/x86_64/vmlinuz
                    append initrd=boot/centos/10-stream/x86_64/initrd.img net.ifnames=0 biosdevname=0 ip=eth0:dhcp inst.ks=http://admin-2.ci.centos.org/ks/cody-n27.ci.centos.org-ks.cfg  

      only eth0 should be configured, and same thing from kickstart used for deployment :

       

      # Network information
      network  --hostname=cody-n27.ci.centos.org
      network  --bootproto=static --device=eth0 --gateway=172.19.3.254 --ip=172.19.0.207 --nameserver=172.19.0.253,172.19.0.252 --netmask=255.255.248.0 --ipv6=auto --activate 
       

      Nothing references eth1, so should be disabled but when booting the machine through network for deployment, it gets ip addresses for both nic , fetches kickstart and then loop on "Checking Storage configuration" forever (adding multiple '.' (dots) to show progress while nothing happens at all

      To ensure that it was related to nic and not storage (partition layout or else) I tried multiple combination until I had to manually set `ip:eth1:none` to completely disable network on second network interface

      That's a regression as in the past, only the first declared interface was configured, and same for kickstart

       

       

              anaconda-maint-list anaconda-maint-list
              farrotin@redhat.com Fabian Arrotin
              anaconda-maint-list anaconda-maint-list
              Release Test Team Release Test Team
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

                Created:
                Updated: