Rhel 7-> 8 upgrade

Archives Ansible Project
,
1

Hello all,

Upgrading from Rhel7-> 8 using ansible. In the pre upgrade task I performed few checks like space check and yum update and subscription-manager registration. No issues in the beginning tasks…

But when performing below task post reboot task Shell:
Yum clean all
Subscription-manager unsubscribe —all
Subscription-manager clean

It is giving me no such file /usr/bin/python

Before running the job In RHEl 7 the python version was pointing to

Python2 /usr/bin/python-> /etc/alternatives/python /etc/alternatives/python → /usr/bin/python2

but after the above task is executed somehow it is changing to
/usr/bin/python-> /etc/alternatives/python /etc/alternatives/python → /usr/bin/python3

When I performed manually the task command there is no issue … could you pls suggest…

Regards
PD

2

That is a Rhel7 to RHEL 8 change. The default python in REHL7 is Python 2.7 while RHEL 8 is python3.6. I would also consider moving the subscription tasks to community.general.redhat_subscription module. That is should automatically change python version, but you may need to disconnect and reconnect to update the user environment.

3

Thank you the problem is before updating to Rhel 8 also when it reboots in the last below task (line#99) the alternatives is updating to python3 link which is not present . Do you suggest to use subscription module in line#79 instead of shell?

  1. name: Get root volume size

  2. set_fact:

  3. root_size: “{{ (ansible_mounts |selectattr(‘mount’, ‘equalto’, ‘/’) | list)[0].size_total }}”

    • name: Get root volume free size

  6. set_fact:

  7. root_available: “{{ (ansible_mounts |selectattr(‘mount’, ‘equalto’, ‘/’) | list)[0].size_available }}”

    • name: root filesystem expansion

  10. block:

    • name: Calculte required additional root space

  13. set_fact:

  14. root_reqd: “{{ ((root_size | int + (4294967296 - root_available | int)) / 1024 / 1024) | int }}”

    • name: Resize root filesystem to {{ root_reqd }}MB

  17. lvol:

  18. vg: rootvg

  19. lv: LogVol00

  20. size: “{{ root_reqd }}M”

  21. resizefs: true

  23. when: root_available | int < 4294967296

    • name: Get /var volume size

  26. set_fact:

  27. var_size: “{{ (ansible_mounts |selectattr(‘mount’, ‘equalto’, ‘/var’) | list)[0].size_total }}”

    • name: Get /var volume free size

  30. set_fact:

  31. var_available: “{{ (ansible_mounts |selectattr(‘mount’, ‘equalto’, ‘/var’) | list)[0].size_available }}”

    • name: /var filesystem expansion

  34. block:

    • name: Calculte required additional var space

  37. set_fact:

  38. var_reqd: “{{ ((var_size | int + (5368709120 - var_available | int)) / 1024 / 1024) | int }}”

    • name: Resize /var filesystem to {{ var_reqd }}MB

  41. lvol:

  42. vg: rootvg

  43. lv: LogVol03

  44. size: “{{ var_reqd }}M”

  45. resizefs: true

  47. when: var_available | int < 5368709120

  49. Pre-upgrade tasks

    • name: Remove unwanted packages

  51. yum:

  52. name: “{{ item }}”

  53. state: absent

  54. loop: “{{ remove_rpms }}”

    • name: Remove unwanted data

  57. file:

  58. path: “{{ item }}”

  59. state: absent

  60. loop: “{{ remove_files }}”

    • name: Get contents of /usr/openv/ folder

  63. find:

  64. paths: /usr/openv/

  65. file_type: any

  66. register: openv

    • name: Remove contents of /usr/openv/ folder

  69. file:

  70. path: “{{ item.path }}”

  71. state: absent

  72. loop: “{{ openv.files }}”

  73. loop_control:

  74. label: “{{ item.path }}”

    • name: Subscribe to RHEL7 Repo

  77. shell: |

  78. yum clean all

  79. subscription-manager register --org=“Xxxxx” --activationkey=“rhel7-prod” --force

    • name: Yum update

  82. yum:

  83. state: latest

    • name: Backup /etc/fstab file

  86. copy:

  87. src: /etc/fstab

  88. dest: /etc/fstab.PreUpgrade

  89. remote_src: yes

  90. force: false

    • name: Remove non system entries from /etc/fstab

  93. shell: |

  94. cat /etc/fstab|egrep -w ‘/ |/opt |/boot |/tmp |/var |swap |^#’ > /tmp/fstab

  95. /bin/cp /tmp/fstab /etc/fstab

    • debug: var=ansible_python_version
    • name: Reboot server

  100. reboot:

  101. reboot_timeout: 3600

    • debug: var=ansible_python_version

  105. Leapp pre-upgrade tasks

    • name: Unsubscribe from satellite

  107. shell: |

  108. yum clean all

  109. subscription-manager unsubscribe --all

  110. subscription-manager clean

Regards

4

Lines 77-79 and 107-110 would be the ones I would change. Using Modules is cleaner, and more predictable typically.

Have you confirmed the user specific settings for the ansible user on the remote host? The per user alternative may be different than the system.

5

Thank you… I ll try to use the modules instead.
And we run ansible with the root user …
also when I rebooting my server manually the default python link is getting updated to pyhon3

Now I raised the issue to RH let’s see what comes up

Thanks again

6

Well, yes. A lot of system tools, especially rpm itself, got updated
to use python3 and the updated /usr/bin/python link because python2 is
*obsolete* and no longer supported. Frankly, by the time you work out
trying nto do the update in place, you could make a full backup of the
whole OS to some separate location, done a clean installation, and
recovered necessary configurations from the backup.

Doing that kind of remotely managed update as a remotely triggered
operation is.... pretty dangerous. There are a lot of ways around it.
My favorite, from long before ansible, was to pick an underused
partition, copy everything I might care about into *that*, and
completely re-install the OS with optional re-partitioning and file
system creation with newer tools along the way.