Hello,

In order to improve the documentation of the automation page, we should add a new section explaining how to use the formulas in a generic way. I have the next in my mind.

Create a new section after all of the formulas are listed. Something like How to use the formulas.
This section would have the next information:

The salt formulas can be used with 2 different approaches: salt master/minion or only salt minion execution.

For the master/minion approach, all of the next steps must be executed in the master machines. If only salt minion option is used, the steps must be executed in all of the minions where the formulas are going to executed.

1. Install the formulas
   Install the formulas that are going to be applied using zypper

2. Create the pillar files structure
   The pillar files work as configuration files for the salt formulas, they are the input the user has to give. In order to use the files, the next structure is the most appropriate one (this is only an example):

a. Create the /srv/pillar folder (this might be already in place)
b. Create a top.sls file in /srv/pillar with the next content. This will apply the hana, netweaver, drbd and ha formulas in the nodes specified by the hostname.

base:
    'hana01,hana02':
        - hana.hana
        - hana.cluster

    'netweaver01,netweaver02':
        - netweaver.netweaver
        - netweaver.cluster

    'drbd01,drbd02':
        - drbd.drbd
        - drbd.cluster

c. Create the folders /srv/pillar/hana, /srv/pillar/netweaver and /srv/pillar/drbd for the pillar files (notice that these names match with the names used in the top.sls file.

d. Add the pillar files in the correct folder. hana folder will have hana.sls and cluster.sls, netweaver folder will have netweaver.sls and cluster.sls, and drbd folder will have drbd.sls and cluster.sls.

Find here an example of the contents you can use for each of the formulas:
hana.sls
netwaever.sls
drbd.sls
cluster.sls

The content of the pillar files must be configured depending on the needed configuration of each of the formulas (notice that in this example, 3 different cluster.sls are used, and each of them might be different, as the HA clusters requirements might be different too.

e. Create a top.sls in /srv/salt folder. It follows the same rules as the pillar top.sls file. Here an example (in this case the folder part is removed, as we are pointing to the formulas with their name):

base:
    'hana01,hana02':
        - hana
        - cluster

    'netweaver01,netweaver02':
        - netweaver
        - cluster

    'drbd01,drbd02':
        - drbd
        - cluster

3. Execute salt for each of the approaches (salt supports more options than those, this is just an example)
   a. For the master/minion approach run:

salt '*' state.highstate

b. For the only minion approach run:

salt-call --local state.highstate

Edit toms:
Related to bsc#1174530

Trento Server installation must be updated as follows:

Step 2 (installing as user root)

Replace
curl -sfL https://get.k3s.io | sh
with
curl -sfL https://get.k3s.io | INSTALL_K3S_SKIP_SELINUX_RPM=true sh

Step 2 (installing as non-root user)
Replace
curl -sfL https://get.k3s.io | sh -s - --write-kubeconfig-mode 644
with
curl -sfL https://get.k3s.io | INSTALL_K3S_SKIP_SELINUX_RPM=true sh -s - --write-kubeconfig-mode 644

This change is to prevent that users running SLES 15 SP3 run into an issue when installing K3s due to missing container-selinux package.

A new version of Trento premium is available (v0.8.1).

The installation process for the Agent hasn't changed. But the installation process for the server must be adjusted to accommodate the new version:

Please replace:

HELM_EXPERIMENTAL_OCI=1 helm upgrade --install \
   TRENTO_SERVER_HOSTNAME \
   oci://registry.suse.com/trento/trento-server \
   --version 0.2.5 \
   --set trento-runner.image.tag=0.7.1 \
   --set trento-web.image.tag=0.7.1 \
   --set-file trento-runner.privateKey=PRIVATE_SSH_KEY 

with

HELM_EXPERIMENTAL_OCI=1 helm upgrade --install \
   trento-server \
   oci://registry.suse.com/trento/trento-server \
   --version 0.3.5 \
   --set-file trento-runner.privateKey=PRIVATE_SSH_KEY

The update sections in the Trento Premium documentation still reflect the update process for the community / open-source version. They must be updated and aligned with the Trento Premium installation process.

Particularly:

• In the case of the Trento Server, the update process should consist of two steps:
  Step 1: export KUBECONFIG=/etc/rancher/k3s/k3s.yaml
  Step 2: HELM_EXPERIMENTAL_OCI=1 helm upgrade --install trento-server oci://registry.suse.com/trento/trento-server --version 0.3.5 --set-file trento-runner.privateKey=PRIVATE_SSH_KEY
• In the case of the Trento Agent, the update process is identical to the installation process.

In art-sol-automation for 15 SP2, there are the following b0rk links:

• Setting up a SAP HANA SR Cost Optimized Infrastructure (SLES-SAP 12 SP1) 404 https://www.suse.com/media/white-paper/sap_hana_sr_cost_optimized_scenario_12_sp1.pdf
• SAP NetWeaver High Availability Cluster 7.40 - Setup Guide (Code 12) 404 https://www.suse.com/media/white-paper/sap_netweaver_availability_cluster_740_setup_guide.pdf
• Simple Stack - SAP NetWeaver High Availability on SUSE Linux Enterprise 12 (SP1 or newer) 404 https://www.suse.com/media/white-paper/sap_netweaver_availability_cluster_740_setup_guide.pdf
• Setting up a SAP HANA SR Cost Optimized Infrastructure (SLES for SAP 12 SP1) 404 https://www.suse.com/media/white-paper/sap_hana_sr_cost_optimized_scenario_12_sp1.pdf

All of these are probably under docs.suse.com/sbp/ now.

The section title DRBD/NFS Automated Configuration in the automation documentation page is not really accurate.

The drbd-formula only configures DRBD, not NFS. In order to use NFS, we need to do things on top of it.

The thing is that SUSE recommends to use DRBD with HA to have a high available NFS share. For that, we configure the HA cluster in top of DRBD to create the NFS exports. But this is not anything the drbd-formula does.

Maybe the title is confusing.

If we want to document how to have high available NFS with DRBD, we will need to add more things (explain that we need to use drbd-formula and habootstrap-formula, and after that apply a specific configuration to the cluster to setup as NFS).

#41 (comment)