Bug 2137404

Hi Mike,
Sharing RHCS Program notes from Jun 29th.
Cephadm-ansible packages were dropped from RHEL9 -

https://docs.google.com/document/d/1aJ8Sbv89VJL6llEDCtxYUO0FBNV3WOyFnWRwRzKDHFU/edit

************************
RHEL 9 Package inclusion for Tools repo : https://bugzilla.redhat.com/show_bug.cgi?id=2100503
Are we shipping cephadm-ansible (for RHEL 9) in 5.2 or 6.0? 
[F2] no, cephadm-ansible is not relevant for OpenStack
************************

Hi George,
Looks like 3.4 to 3.9 does not apply to RHEL9 and required manual steps to prepare nodes for deployment on RHEL9.
Will get the documentation updated.

Vivek - can we have Doc changes tracked for RHEL9 support for RHCS 5.x [RHEL 8 containers]?


Regards,
Veera

(In reply to Veera Raghava Reddy from comment #8)
> Hi George,
> Looks like 3.4 to 3.9 does not apply to RHEL9 and required manual steps to
> prepare nodes for deployment on RHEL9.
> Will get the documentation updated.
> 
> Vivek - can we have Doc changes tracked for RHEL9 support for RHCS 5.x [RHEL
> 8 containers]?
> 
> 
> Regards,
> Veera

ACK Veera, I will have the doc bug created for tracking the same. Until them I am keeping the needinfo intact

Regards,
Vivek Das

@vdas - can you provide an outline of the manual steps so that I can at least provide something back to the customer while we're waiting on the docs to be updated?

@mhackett - FYI - this seems to have stalled
No doc updates for the RHEL9 install

Review the comments from George and take to closure on the following with Doc team
- Add note to Skip cephadm-ansible steps for 5.x on RHEL 9
- Add manual pre-deployment steps

Please find below steps.


Scenario 1: RHCS 5.x host re-provisioning from RHEL 8 to RHEL 9.
===============================================================
Version : 16.2.10-69.el8cp 
Status : Verification completed
Host re-provisioning  : RHEL 8 to RHEL 9

Steps:
=====
1. Deploy RHCS 5 on RHEL 8 with necessary hosts and daemons.
2. Add a second admin node in the cluster (To manage the cluster when admin node re-provisioned).
3. Set the  “noout” flag.
4. Drain one host.
5. Removed the host.
6. Re-provision respective host to RHEL 9.
7. After host upgrade,Install regular dependencies (lvm2, podman, chrony, python3) on the host.
Note : Skip cephadm-ansible steps for 5.x on RHEL 9, Add manually above dependencies.
8. Add “ceph.pub” key on the re-provisioned node.
9. Add host again to the cluster.
10. Verify the daemons on re-provision host running successfully with the same ceph version.
11. Repeat same with all other hosts as well (Step 3 to step 9)
12. Take backup of ceph binary (/usr/sbin/cephadm) and ceph.pub(/etc/ceph) from the first admin node.
13. Follow the same approach to re-provision admin nodes and meantime use a second admin node to manage clusters.
14. Add the backup files again to the node.
15. Add admin nodes again to cluster using the second admin node.
16. Follow the same approach to re-provision second admin nodes and use a first admin node to manage clusters.
17. Unset the  “noout” flag.
18. Verify ceph version,cluster status is healthy with all demons working as expected after re-provision.

Hi @George Law

For rhcs 5 installation in rhel 9 we need to skip cephadm-ansible installation since as of now cephadm-ansible not support rhel 9 for rhcs 5.

We need to manually install dependencies (lvm2, podman, chrony, python3).

Apart for cephadm-ansible installation and running playbook other steps remain same.


https://access.redhat.com/documentation/en-us/red_hat_ceph_storage/5/html-single/installation_guide/index#bootstrapping-a-new-storage-cluster_install

Hi,

Changes looks good.
Moving bug to "verified"

Thanks,
Mohit Bisht

Hi George,
The decision to not generate packages for RHEL 9 were based on the effort and priorities required to release 5. x containers support for OSP17. But see how to best convey with customers. Maybe Fedrico can help to position it appropriately.