TripleO deployment fails using Centos7 + stable/liberty

[Carlos Camacho]

After a manual deployment of Tripleo using stable/liberty faced an issue in the NetworkDeployment step when deploying the overcloud.
After further research the problem is due to the fact that the baremetal nodes are not even starting up.

The nodes are starting, then they do "something" not visible, and after a reboot is shown this error which cause the deployment error
----------------------------------------------------------
"error: not a correct XFS inode."
"error: file '/boot/grub2/i386-pc/normal.mod' not found."
----------------------------------------------------------

How to reproduce:
  Follow the steps in:
  https://gist.github.com/ccamacho/aa136fd87a23b4898185b8c95081905c

This might not be hitting CI as EXT4 is configured instead XFS https://review.openstack.org/#/c/318400/ and there is only
the master branch..

Workarround:
  export FS_TYPE=ext4
  rm -rf overcloud-full.*
  glance image-delete overcloud-full
  glance image-delete overcloud-full-initrd
  glance image-delete overcloud-full-vmlinuz
  ./tripleo-ci/scripts/tripleo.sh --overcloud-image

Im still having issues with the deployment but, at least I can log in as the heat-admin user in the controller.

[Carlos Camacho (ccamacho)]

This is the last error

[Carlos Camacho (ccamacho)]

Using a prebuilt image..:
#From stack home
source stackrc
rm -rf overcloud-full.*
glance image-delete overcloud-full
glance image-delete overcloud-full-initrd
glance image-delete overcloud-full-vmlinuz
wget http://artifacts.ci.centos.org/artifacts/rdo/images/liberty/delorean/stable/overcloud-full.tar
tar -xvf overcloud-full.tar
./tripleo-ci/scripts/tripleo.sh --overcloud-images
openstack overcloud deploy --libvirt-type qemu --ntp-server pool.ntp.org --templates /home/stack/tripleo-heat-templates -e /home/stack/tripleo-heat-templates/overcloud-resource-registry-puppet.yaml

Now, I can see the novacompute node but not the controller...

[Carlos Camacho (ccamacho)]

Using the prev image, I just have re-executed the deployment, now both nodes with errors booting up.

[Carlos Camacho (ccamacho)]

Now Im getting hit by this issue: https://bugs.launchpad.net/tripleo/+bug/1546749

[Carlos Camacho (ccamacho)]

Fix proposed: https://review.openstack.org/#/c/327773/

[David Hill (david-hill-ubisoft)]

Hello guys,

  I'm having this issue right now with FS_TYPE=xfs and can confirm that using ext4 instead seems to improve the situation. The problem I'm facing with your workaround is that it's a workaround and not solving the XFS issues. Sometimes it works and some times it doesn't. What could cause this intermittent behavior?

Thank you very much,

David Hill