osism
insufficient ansible version in Ubuntu 22.04 image
I'm running the testbed Deployment against regiocloud.
On the Step
make ENVIRONMENT=regiocloud manager
I get: ` [...] TASK [Point out that the following task takes some time and does not give any output] ************************************************************************************************ ok: [localhost] => { "msg": "The task 'Run manager part 0' runs an Ansible playbook on the manager. There is no further output of this here. It takes a few minutes for this task to complete." }
TASK [Run manager part 0] *********************************************************************************************************************************************************** [...] ***************************************************", "ok: [testbed-manager]", "", "TASK [Install netaddr in venv] *************************************************", "ok: [testbed-manager]", "", "TASK [Install ansible-core in venv] ********************************************", "fatal: [testbed-manager]: FAILED! => {"changed": false, "cmd": ["/opt/venv/bin/pip3", "install", "ansible-core<2.19.0,>=2.18.0"], "msg": "\n:stderr: ERROR: Could not find a version that satisfies the requirement ansible-core<2.19.0,>=2.18.0 (from versions: 0.0.1a1, 2.11.0b1, 2.11.0b2, 2.11.0b3, 2.11.0b4, 2.11.0rc1, 2.11.0rc2, 2.11.0, 2.11.1rc1, 2.11.1, 2.11.2rc1, 2.11.2, 2.11.3rc1, 2.11.3, 2.11.4rc1, 2.11.4, 2.11.5rc1, 2.11.5, 2.11.6rc1, 2.11.6, 2.11.7rc1, 2.11.7, 2.11.8rc1, 2.11.8, 2.11.9rc1, 2.11.9, 2.11.10rc1, 2.11.10, 2.11.11rc1, 2.11.11, 2.11.12rc1, 2.11.12, 2.12.0b1, 2.12.0b2, 2.12.0rc1, 2.12.0, 2.12.1rc1, 2.12.1, 2.12.2rc1, 2.12.2, 2.12.3rc1, 2.12.3, 2.12.4rc1, 2.12.4, 2.12.5rc1, 2.12.5, 2.12.6rc1, 2.12.6, 2.12.7rc1, 2.12.7, 2.12.8rc1, 2.12.8, 2.12.9rc1, 2.12.9, 2.12.10rc1, 2.12.10, 2.13.0b0, 2.13.0b1, 2.13.0rc1, 2.13.0, 2.13.1rc1, 2.13.1, 2.13.2rc1, 2.13.2, 2.13.3rc1, 2.13.3, 2.13.4rc1, 2.13.4, 2.13.5rc1, 2.13.5, 2.13.6rc1, 2.13.6, 2.13.7rc1, 2.13.7, 2.13.8rc1, 2.13.8, 2.13.9rc1, 2.13.9, 2.13.10rc1, 2.13.10, 2.13.11rc1, 2.13.11, 2.13.12rc1, 2.13.12, 2.13.13rc1, 2.13.13, 2.14.0b1, 2.14.0b2, 2.14.0b3, 2.14.0rc1, 2.14.0rc1.post0, 2.14.0rc2, 2.14.0, 2.14.1rc1, 2.14.1, 2.14.2rc1, 2.14.2, 2.14.3rc1, 2.14.3, 2.14.4rc1, 2.14.4, 2.14.5rc1, 2.14.5, 2.14.6rc1, 2.14.6, 2.14.7rc1, 2.14.7, 2.14.8rc1, 2.14.8, 2.14.9rc1, 2.14.9, 2.14.10rc1, 2.14.10, 2.14.11rc1, 2.14.11, 2.14.12rc1, 2.14.12, 2.14.13, 2.14.14rc1, 2.14.14, 2.14.15rc1, 2.14.15, 2.14.16rc1, 2.14.16, 2.14.17rc1, 2.14.17, 2.14.18rc1, 2.14.18, 2.15.0b1, 2.15.0b2, 2.15.0b3, 2.15.0rc1, 2.15.0rc2, 2.15.0, 2.15.1rc1, 2.15.1, 2.15.2rc1, 2.15.2, 2.15.3rc1, 2.15.3, 2.15.4rc1, 2.15.4, 2.15.5rc1, 2.15.5, 2.15.6rc1, 2.15.6, 2.15.7rc1, 2.15.7, 2.15.8, 2.15.9rc1, 2.15.9, 2.15.10rc1, 2.15.10, 2.15.11rc1, 2.15.11, 2.15.12rc1, 2.15.12, 2.15.13rc1, 2.15.13, 2.16.0b1, 2.16.0b2, 2.16.0rc1, 2.16.0, 2.16.1rc1, 2.16.1, 2.16.2, 2.16.3rc1, 2.16.3, 2.16.4rc1, 2.16.4, 2.16.5rc1, 2.16.5, 2.16.6, 2.16.7rc1, 2.16.7, 2.16.8rc1, 2.16.8, 2.16.9rc1, 2.16.9, 2.16.10rc1, 2.16.10, 2.16.11rc1, 2.16.11, 2.16.12rc1, 2.16.12, 2.16.13rc1, 2.16.13, 2.16.14rc1, 2.16.14, 2.17.0b1, 2.17.0rc1, 2.17.0rc2, 2.17.0, 2.17.1rc1, 2.17.1, 2.17.2rc1, 2.17.2rc2, 2.17.2, 2.17.3rc1, 2.17.3, 2.17.4rc1, 2.17.4, 2.17.5rc1, 2.17.5, 2.17.6rc1, 2.17.6, 2.17.7rc1, 2.17.7, 2.17.8rc1, 2.17.8, 2.17.9rc1, 2.17.9, 2.17.10rc1, 2.17.10, 2.17.11rc1, 2.17.11, 2.17.12rc1, 2.17.12)\nERROR: No matching distribution found for ansible-core<2.19.0,>=2.18.0\n"}", "", "PLAY RECAP *********************************************************************", "testbed-manager : ok=12 changed=1 unreachable=0 failed=1 skipped=7 rescued=0 ignored=0 "]} `
The image used for the manager was "Ubuntu 22.04" (default value for regiocloud.
I "fixed" this by setting the image to "Ubuntu 24.04" (I'm not sure, but I guess OSISM 8.0.1 isn't officially comptible with Ubuntu 24.04. So I expect other problems to come up down the line. BTW: Is there a compatibility matrix somewhere?)
Not sure what the correct fix is here:
- Add an "apt update, apt upgrade" step in the playbook before that venv is created on the manager?
- Add a "pip install -U pip" step before installing ansible? (Honestly I didn't figure how to connect to the manager vm. The key that I found in ./terraform/.terraform_state didn't work on the ubuntu or dragon user. Otherwise I would have tested 1. and 2. myself.)
Any pointers on how to fix this properly and I'll come up with a Merge Request myself.
You have to use Ubuntu 24.04. It probably makes sense to cleanup the provided environment files in osism/testbed. At least we should use Ubuntu 24.04 as default for Regiocloud.
Ubuntu 24.04 is only necessary for the Seed node to deploy the Manager node the first time (because we have to install latest Ansible on this node which requires Ubuntu 24.04). Ubuntu 22.04 should still work for all other nodes. But it‘s recommnded to use Ubuntu 24.04.
I can do that and make a pull request. Should I also change the volume type default for regiocloud from "ssd" to "__DEFAULT__"? At least I had to do that because "ssd" wasn't available in my project. I have no problem with that, since I just need an Openstack API to run administrator privilege requests against. And I guess for a testbed __DEFAULT__ (I guess Ceph HDD) is good enough.
I fixed the volume types on your project. Please use ssd. This speeds up the deployment. The hdd backed Ceph cluster (DEFAULT) is a little bit slow.
Please change default image for Regiocloud and default profiles to Ubuntu 24.04.