Ansible fails on Debian 10 with ceph_repository_type: distro (stable-4.0)

[RubenGarcia]

Bug Report

What happened:

Ansible fails with error when running the line
ceph-volume --cluster ceph lvm batch --bluestore DEVICES --report --format=json

When running the line in the command line, I get
--> DEPRECATION NOTICE
--> You are using the legacy automatic disk sorting behavior
--> The Pacific release will change the default to --no-auto
--> passed data devices: 0 physical, 180 LVM
--> relative data size: 1.0
--> IndexError: list index out of range

What you expected to happen:

Correct installation

How to reproduce it (minimal and precise):

Install debian 10.
Enable backports and install ceph with

apt install ceph/buster-backports ceph-mgr/buster-backports ceph-base/buster-backports ceph-osd/buster-backports ceph-mon/buster-backports ceph-common/buster-backports libradosstriper1/buster-backports smartmontools/buster-backports

Setup branch to stable-4.0
Use ceph_repository_type: distro
Run the ansible

Share your group_vars files, inventory and full ceph-ansibe log

Environment:

• OS (e.g. from /etc/os-release): Debian GNU/Linux 10 (buster)
• Kernel (e.g. uname -a): Linux lexis-expstorage-2 4.19.0-17-amd64 Add Ceph Playbook #1 SMP Debian 4.19.194-3 (2021-07-18) x86_64 GNU/Linux
• Docker version if applicable (e.g. docker version): N/A
• Ansible version (e.g. ansible-playbook --version): 2.9.26
• ceph-ansible version (e.g. git head or tag or stable branch): stable-4.0
• Ceph version (e.g. ceph -v): ceph version 14.2.21 (5ef401921d7a88aea18ec7558f7f9374ebd8f5a6) nautilus (stable)

Possible solutions:
-Indicate in the documentation that this configuration is not working at the moment.
-Ask Debian to update their package to the version of ceph which no longer has the bug.

[RubenGarcia]

group_vars.zip
ansiblelog.zip
hosts.txt

[RubenGarcia]

The problem was that the devices were multipath.
I recommend that a better error message like

No supported disks found. Only raw disks and lvm volumes are supported (e.g. no raw multipath).

is added in a try_catch clause at the proper location.

[guits]

@RubenGarcia At first glance, yes, there's a bug anyway here, it should throw a 'nice' error message rather than a meaningless error from user perspective.

By the way, are the 180 devices reported in the following output well LVM devices?

--> passed data devices: 0 physical, 180 LVM

[RubenGarcia]

These 180 lvm devices were created using this code:

cd /dev/mapper
for i in mpath*; do pvcreate $i; done
for i in mpath*; do vgcreate $i /dev/mapper/$i; done
for i in mpath*; do lvcreate -l 100%FREE -n $i $i; done

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in a week if no further activity occurs. Thank you for your contributions.

[RubenGarcia]

Seriously, you need to configure your github-actions to avoid closing issues which are not fixed.

[guits]

@RubenGarcia this isn't a ceph-ansible issue.
Do you mind opening a tracker at https://tracker.ceph.com/projects/ceph-volume ? Thanks!