All,
I notice that most of our Fedora Server documentation describes using cockpit and command-line and graphical apps.
While I have no objection to this and I believe it works well for maintaining a small number of servers and/or services, it is my belief this becomes quite cumbersome for more than a handful of servers and/or services.
I have been asked to edit the NFS Server service installation and setup documentation currently found in the Staging Environment. The docs should cover both server side and client side procedures. I am more that willing to edit the procedures prescribed in that document. But I prefer using ansible to maintain my server environment. It becomes permanent documentation (in case I am no longer available) and it becomes repeatable by only modifying the "environment files" if a server requires a rebuild.
I am looking for the opinion and experience of the community in deciding whether or not we should document both types of procedures.
John, In my experience, mature environments are migrating to Ansible'ized documentation maintanined in GitLab.It'll be interesting to see what others can contribute. Thank you.R,-Joe
On Thursday, May 2, 2024 at 06:07:09 PM EDT, John W. Himpel john@jlhimpel.net wrote: All,
I notice that most of our Fedora Server documentation describes using cockpit and command-line and graphical apps.
While I have no objection to this and I believe it works well for maintaining a small number of servers and/or services, it is my belief this becomes quite cumbersome for more than a handful of servers and/or services.
I have been asked to edit the NFS Server service installation and setup documentation currently found in the Staging Environment. The docs should cover both server side and client side procedures. I am more that willing to edit the procedures prescribed in that document. But I prefer using ansible to maintain my server environment. It becomes permanent documentation (in case I am no longer available) and it becomes repeatable by only modifying the "environment files" if a server requires a rebuild.
I am looking for the opinion and experience of the community in deciding whether or not we should document both types of procedures.
-- _______________________________________________ server mailing list -- server@lists.fedoraproject.org To unsubscribe send an email to server-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/server@lists.fedoraproject.org Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
Am 03.05.2024 um 00:06 schrieb John W. Himpel john@jlhimpel.net:
All,
I notice that most of our Fedora Server documentation describes using cockpit and command-line and graphical apps.
While I have no objection to this and I believe it works well for maintaining a small number of servers and/or services, it is my belief this becomes quite cumbersome for more than a handful of servers and/or services.
I have been asked to edit the NFS Server service installation and setup documentation currently found in the Staging Environment. The docs should cover both server side and client side procedures. I am more that willing to edit the procedures prescribed in that document. But I prefer using ansible to maintain my server environment. It becomes permanent documentation (in case I am no longer available) and it becomes repeatable by only modifying the "environment files" if a server requires a rebuild.
I am looking for the opinion and experience of the community in deciding whether or not we should document both types of procedures.
I strongly agree with adding Ansible procedures to our documentation. We have committed to Ansible in our technical specifications (https://docs.fedoraproject.org/en-US/server-working-group/docs/server-techni...) Our Server Roles are Ansible supported services. And it is the core of our project "Using Ansible to install and configure Wildfly“ (issue #60, https://pagure.io/fedora-server/issue/60) We chose Wildfly als our PoC project. Unfortunately it turned out, that Wildfly itself provides so many quirks which we had to master, that we didn’t came to Ansible until now.
Maybe it is a good idea to use the NFS case as an opportunity to use this as a PoC for our Server Roles. NFS is obviously much easier to master and we can concentrate on the Ansible part instead on the quirks of the software itself.
And a next good candidate would be our „Post-install Tasks“, I think (although this is of course not a „Server Role“).
Regarding the current NFS article:
I think we should leave the first part as it is until "Preparation" and insert a new part before Preparation: The NFS Server Role. And describe in this part to use Ansible for setup. And than continue with a new sub-header like „Manual Setup“ and describe the manual procedure.
Or we can split it into separate Sub-docs like
== The Fedora Server NFS Server Role With some general information a === Setting up NFS Server Role via Ansible === Setting up NFS Server Role via CLI === Setting up a Fedora Linux NFS client (desktop an server)
We have to discard the Cockpit part about NFS server, because the module was never really developed further and is now orphaned, if I remember correctly.
And then we have to discuss and decide how to maintain and distribute the Ansible files.
-- Peter Boy https://fedoraproject.org/wiki/User:Pboy PBoy@fedoraproject.org
Timezone: CET (UTC+1) / CEST (UTC+2)
Fedora Server Edition Working Group member Fedora Docs team contributor and board member Java developer and enthusiast
clustershell, which is what I use is great for administration of large numbers of machines.
In case you are not familiar with that method it is super simple. You create a policy in PAM/SUDO and SSH key common to all machines, say 6000 of them.
You can then use cluster shell to copy a script to all 6000 machines, and then execute it, say for updating the time and date for example, checking the date time, installing RPMS....etc
You just put what commands associated with the key into what login ID you want to restrict.
But I don't really see a need for something like cockpit and never use it as it is too simplistic, like I am using a Microsoft product.
:-)
On Thu, May 2, 2024 at 3:07 PM John W. Himpel john@jlhimpel.net wrote:
All,
I notice that most of our Fedora Server documentation describes using cockpit and command-line and graphical apps.
While I have no objection to this and I believe it works well for maintaining a small number of servers and/or services, it is my belief this becomes quite cumbersome for more than a handful of servers and/or services.
I have been asked to edit the NFS Server service installation and setup documentation currently found in the Staging Environment. The docs should cover both server side and client side procedures. I am more that willing to edit the procedures prescribed in that document. But I prefer using ansible to maintain my server environment. It becomes permanent documentation (in case I am no longer available) and it becomes repeatable by only modifying the "environment files" if a server requires a rebuild.
I am looking for the opinion and experience of the community in deciding whether or not we should document both types of procedures.
-- _______________________________________________ server mailing list -- server@lists.fedoraproject.org To unsubscribe send an email to server-leave@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/server@lists.fedoraproject.org Do not reply to spam, report it: https://pagure.io/fedora-infrastructure/new_issue
Hi John,
"John W. Himpel" john@jlhimpel.net writes:
All,
I notice that most of our Fedora Server documentation describes using cockpit and command-line and graphical apps.
While I have no objection to this and I believe it works well for maintaining a small number of servers and/or services, it is my belief this becomes quite cumbersome for more than a handful of servers and/or services.
I have been asked to edit the NFS Server service installation and setup documentation currently found in the Staging Environment. The docs should cover both server side and client side procedures. I am more that willing to edit the procedures prescribed in that document. But I prefer using ansible to maintain my server environment. It becomes permanent documentation (in case I am no longer available) and it becomes repeatable by only modifying the "environment files" if a server requires a rebuild.
I am looking for the opinion and experience of the community in deciding whether or not we should document both types of procedures.
I agree and I guess every admin made the transition from ssh-ing into servers to clusterssh to Ansible/Chef/Salt/etc.
However, I think we should make a distinction between deploying a configuration and the configuration itself. These are two separate processes and in the end it shouldn't matter *how* we push the configuration to the nfs server & client, it's the config that matters. How we push the config is a separate topic, that the documentation can also tackle, but in a separate chapter.
Cheers,
Dan
Am 03.05.2024 um 07:54 schrieb Dan Čermák dan.cermak@cgc-instruments.com:
. . . it shouldn't matter *how* we push the configuration to the nfs server & client, it's the config that matters. How we push the config is a separate topic, that the documentation can also tackle, but in a separate chapter.
100% agreement. You've hit it / formulated it better than I did myself, which is my intention / concept I'm pursuing with the documentation.
-- Peter Boy https://fedoraproject.org/wiki/User:Pboy PBoy@fedoraproject.org
Timezone: CET (UTC+1) / CEST (UTC+2)
Fedora Server Edition Working Group member Fedora Docs team contributor and board member Java developer and enthusiast
On Fri, May 03, 2024 at 10:29:39AM GMT, Peter Boy wrote:
Am 03.05.2024 um 07:54 schrieb Dan Čermák dan.cermak@cgc-instruments.com:
. . . it shouldn't matter *how* we push the configuration to the nfs server & client, it's the config that matters. How we push the config is a separate topic, that the documentation can also tackle, but in a separate chapter.
100% agreement. You've hit it / formulated it better than I did myself, which is my intention / concept I'm pursuing with the documentation.
Yeah, thats a good way to put it. :)
The issue I see here (and it's one I have struggled with over the years) is that if you try and say "Here, we provide and maintain an ansible script to setup NFS", your script will not handle many of the cases people have, and you are basically duplicating ansible galaxy/collections and committing to a LOT of maint.
On the other hand, if you don't provide anything, you are just a 'assemble your own server from our high quality blocks' which isn't what Fedora editions are supposed to be. :)
I don't have the answer, just thought I would mention the problem again.
kevin
* John W. Himpel [02/05/2024 17:06] :
But I
prefer using ansible to maintain my server environment. It becomes permanent documentation (in case I am no longer available) and it becomes repeatable by only modifying the "environment files" if a server requires a rebuild.
Over the years, this has become the primary reason I use Ansible. It functions as both code and documentation, making it impossible to change one without changing the other.
I am looking for the opinion and experience of the community in deciding whether or not we should document both types of procedures.
I suspect that our target audience are people who straddle the line between maintaining a handful of servers and having a whole cluster of them. This would suggest that it would be useful to have all three documentation types (GUI, CLI, automated).
Like others, I think writing the configuration and deploying the configuration are two seperate steps. I also think that creating an Ansible playbook is trivial to do once you've actually written the CLI steps and that showing this process is more useful than publishing a playbook.
Emmanuel
server@lists.fedoraproject.org