Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Bare metal deployment guide #222

Merged
merged 8 commits into from
Jun 18, 2019
Merged

Bare metal deployment guide #222

merged 8 commits into from
Jun 18, 2019

Conversation

Klaven
Copy link

@Klaven Klaven commented May 20, 2019

@r0ckarong I'm struggling here. for some reason the references are not working. Also, I refereed to it as skuba, not sure if I should be doing that or caaspctl. I feel like the one section about where/how to get the autoyast file should go away and we should just link the example when the code is open sourced.

@Klaven Klaven added the Blocked Blocked by lack of information or external factors label May 20, 2019
@Klaven Klaven changed the base branch from develop to adoc May 20, 2019 20:54
@Klaven Klaven requested a review from r0ckarong May 20, 2019 20:56
@r0ckarong r0ckarong added DeploymentGuide Fix will change the Deployment Guide On-Hold PR needs work and should not be merged even if reviews are complete v4 CaaSP v4 labels May 21, 2019
@r0ckarong
Copy link
Contributor

@Klaven could you please allow maintainer edits for this pull request:
https://help.github.com/en/articles/allowing-changes-to-a-pull-request-branch-created-from-a-fork

This makes work a lot easier. Thanks

@r0ckarong r0ckarong changed the base branch from adoc to master May 22, 2019 10:40
@stefsuse stefsuse requested a review from nkoranova May 24, 2019 01:02
Copy link
Member

@flavio flavio left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, but I think it should be extended to mention:

  • The node will have only a root user enable, with no password set due to security reasons
  • The user must edit the template and include his own ssh keys (maybe show a snippet with multiple identities added)
  • Point out the only way to connect to the node, in the beginning, is to ssh as root
  • Point out the root user can be enabled after the 1st login by using the passwd command
  • Explain how to perform the registration against a SMT/RMT server. What are the values that have to be entered. The point is, the snippet we generate is for SCC, how different is it to connect to a SMT server/RMT?

As always, we don't have to rewrite the existing documentation of SLES. If there's already some existing documentation let's just point to it (for example: the registration against SMT/RMT should already be part of some manual)

@flavio
Copy link
Member

flavio commented May 24, 2019

I looks like @dcoopr has already spotted this issue and proposed the same remediation I was suggesting: https://bugzilla.suse.com/show_bug.cgi?id=1135998

@r0ckarong r0ckarong removed Blocked Blocked by lack of information or external factors On-Hold PR needs work and should not be merged even if reviews are complete labels May 27, 2019
@r0ckarong r0ckarong added this to the Sprint 6 milestone May 27, 2019
@r0ckarong r0ckarong self-assigned this May 27, 2019
@Klaven Klaven force-pushed the bare-metal branch 2 times, most recently from 70f4146 to 443bc88 Compare May 29, 2019 13:22
@Klaven Klaven changed the title [WIP] Bare metal Bare metal deployment guide May 29, 2019
Copy link
Member

@flavio flavio left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM

@jordimassaguerpla
Copy link
Member

You may need to review the docs based on the result of this SUSE/skuba#279

@Klaven
Copy link
Author

Klaven commented May 30, 2019

@jordimassaguerpla the person who completes #279 should update the documentation once it is complete. nothing to hold this PR for.

Includes documentation on where to find an example autoyast file as
well as links to the other required documentation.
Copy link
Contributor

@thehejik thehejik left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thanks - I will provide another PR once SUSE/skuba#283 will be enhanced by caasp-node pattern installation and merged.

Not needed anymore - Markus uses general "placeholder" word and also added ntp_server note.

@r0ckarong r0ckarong modified the milestones: Sprint 6, Sprint 7 Jun 3, 2019
fixes SUSE/avant-garde#389

Signed-off-by: Jordi Massaguer Pla <jmassaguerpla@suse.de>
@jordimassaguerpla
Copy link
Member

@r0ckarong
Copy link
Contributor

@Klaven I'm missing several things in the current version:

  1. The AutoYaST example file sets up only a node but not a master. How does one set up a master? Which pattern needs to be changed?
  2. Will this actually work like this? The "install required packages" comes before the "register the repositories holding the package" in the file. Is this parsed out of order?
  3. How do you define the size of the cluster with this method?
  4. Do you simply repeat these steps for as many nodes as you need?

@thehejik
Copy link
Contributor

I'm involved in autoyast as well so I will try to answer...

@Klaven I'm missing several things in the current version:

1. The AutoYaST example file sets up only a node but not a master. How does one set up a master? Which pattern needs to be changed?

There is no difference, master and worker nodes are using the very same pattern

2. Will this actually work like this? The "install required packages" comes before the "register the repositories holding the package" in the file. Is this parsed out of order?

If you are talking about item order in autoyast.xml file it doesn't matter - SLE will handle the item order automatically during installation.

3. How do you define the size of the cluster with this method?
4. Do you simply repeat these steps for as many nodes as you need?

yup, just repeat these steps for as many nodes you need (or how many machines you have)

@r0ckarong
Copy link
Contributor

@thehejik @Klaven Please review the updated version.


The default production scenario consists of 8 nodes:

* 2 Load Balancers
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think that skuba command supports more than one load-balancer.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This is how it's described in the architecture description. Any idea how this would be handled?

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No idea, maybe it can be achievable by using round-robin DNS entry (two lb hosts with the same public IP) but I would describe only a case with 1 LB. After clarifying we can modify again.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I talked to Rafa we need to clarify the load balancer stuff more.


[NOTE]
====
To account for hardware/platform specific setup criteria (legacy BIOS vs. (U)EFI, drive partitioning, networking etc.),
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've verified that current ay file is compatible with BIOS and EFI as well.

@thehejik
Copy link
Contributor

Apart of the unclear LBs count LGTM

@r0ckarong
Copy link
Contributor

The LB topic will be clarified at a later date with: #261.
This PR will be merged.

@r0ckarong r0ckarong merged commit 627c7fe into SUSE:master Jun 18, 2019
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
DeploymentGuide Fix will change the Deployment Guide v4 CaaSP v4
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants