Merge pull request #2138 from cgoveas/main

Updating 1.4.3 changes

docs/source/InstallationGuides/BuildingClusters/OneAPI.rst → docs/source/InstallationGuides/Benchmarks/OneAPI.rst

@@ -1,5 +1,5 @@
-Install oneAPI for MPI jobs
-___________________________
+Install oneAPI for MPI jobs on Intel processors
+________________________________________________
 
 **Pre-requisites**
 

docs/source/InstallationGuides/Benchmarks/OpenMPI_AOCC.rst

@@ -0,0 +1,73 @@
+Open MPI AOCC HPL benchmark for AMD processors
+----------------------------------------------
+
+**Prerequisites**
+
+* Provision the cluster and install slurm on all cluster nodes.
+* OpenMPI should be installed and compiled with slurm on all cluster nodes or should be available on the NFS share.
+
+
+**To execute multi-node jobs**
+
+1. Update the following parameters in ``/etc/slurm/slurm.conf``: ::
+
+    SelectType=select/cons_tres
+    SelectTypeParameters=CR_Core
+    TaskPlugin=task/affinity,task/cgroup
+
+2. Restart ``slurmd.service`` on all compute nodes. ::
+
+    systemctl stop slurmd
+    systemctl start slurmd
+
+3. Once the service restarts on the compute nodes, restart ``slurmctld.service`` on the manager node. ::
+
+        systemctl stop slurmctld.service
+        systemctl start slurmctld.service
+
+4. Job execution can now be initiated. Provide the host list using ``srun`` and ``sbatch``. For example:
+
+For a job to run on multiple nodes (``omnianode00001.omnia.test``,``omnianode00006.omnia.test`` and,``omnianode00005.omnia.test``) and OpenMPI is compiled and installed on the NFS share (``/home/omnia-share/openmpi/bin/mpirun``), the job can be initiated as below: ::
+
+
+    srun -N 3 --partition=mpiexectrial /home/omnia-share/openmpi/bin/mpirun -host omnianode00001.omnia.test,omnianode00006.omnia.test,omnianode00005.omnia.test ./amd-zen-hpl-2023_07_18/xhpl
+
+For a batch job using the same parameters, the command would be: ::
+
+
+
+    #!/bin/bash
+    
+    #SBATCH --job-name=test
+    
+    #SBATCH --output=test.log
+    
+    #SBATCH --partition=normal
+    
+    #SBATCH -N 3
+    
+    #SBATCH --time=10:00
+    
+    #SBATCH --ntasks=2
+    
+     
+    
+     
+    
+    source /home/omnia-share/setenv_AOCC.sh
+    
+    export PATH=$PATH:/home/omnia-share/openmpi/bin
+    
+    export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/home/omnia-share/openmpi/lib
+    
+
+     
+    
+    mpirun  -host omnianode00001.omnia.test,omnianode00005.omnia.test ./amd-zen-hpl-2023_07_18/xhpl
+    
+    srun sleep 30
+
+
+
+
+

docs/source/InstallationGuides/Benchmarks/index.rst

@@ -0,0 +1,6 @@
+Running HPC benchmarks on omnia clusters
+=========================================
+
+.. toctree::
+    OneAPI
+    OpenMPI_AOCC

docs/source/InstallationGuides/BuildingClusters/Authentication.rst

@@ -196,9 +196,9 @@ Once user accounts are created, admins can enable passwordless SSH for users to
     * If ``enable_omnia_nfs`` is true in ``input/omnia_config.yml``, follow the below steps to configure an NFS share on your LDAP server:
         - From the manager node:
             1. Add the LDAP server IP address to ``/etc/exports``.
-            2. Run ``exports -ra`` to enable the NFS configuration.
+            2. Run ``exportfs -ra`` to enable the NFS configuration.
         - From the LDAP server:
-            1. Add the required fstab entries in ``/etc/fstab``.
+            1. Add the required fstab entries in ``/etc/fstab``. (The corresponding entry will be available on the compute nodes in ``/etc/fstab``)
             2. Mount the NFS share using ``mount manager_ip: /home/omnia-share /home/omnia-share``.
     * If ``enable_omnia_nfs`` is false in ``input/omnia_config.yml``, ensure the user-configured NFS share is mounted on the LDAP server.
 

docs/source/InstallationGuides/BuildingClusters/index.rst

@@ -1,13 +1,32 @@
 Configuring the cluster
 =======================
+
+**Features enabled by omnia.yml**
+
+    * Centralized authentication: Once all the required parameters in `security_config.yml <schedulerinputparams.html>`_ are filled in, ``omnia.yml`` can be used to set up FreeIPA/LDAP.
+
+    * Slurm: Once all the required parameters in `omnia_config.yml <schedulerinputparams.html>`_ are filled in, ``omnia.yml`` can be used to set up slurm.
+
+    * Login Node (Additionally secure login node)
+
+    * Kubernetes: Once all the required parameters in `omnia_config.yml <schedulerinputparams.html>`_ are filled in, ``omnia.yml`` can be used to set up kubernetes.
+
+    * BeeGFS bolt on installation: Once all the required parameters in `storage_config.yml <schedulerinputparams.html>`_ are filled in, ``omnia.yml`` can be used to set up NFS.
+
+    * NFS bolt on support: : Once all the required parameters in `storage_config.yml <schedulerinputparams.html>`_ are filled in, ``omnia.yml`` can be used to set up BeeGFS.
+
+
+
 .. toctree::
     schedulerinputparams
     schedulerprereqs
     installscheduler
     Authentication
-    OneAPI
     BeeGFS
     NFS
+    OneAPI
+    OpenMPI_AOCC
     KernelUpdateRHEL
 
 
+

docs/source/InstallationGuides/BuildingClusters/installscheduler.rst

@@ -1,20 +1,27 @@
 Building clusters
 ------------------
 
-1. In the ``input/omnia_config.yml`` file, provide the `required details <schedulerinputparams.html>`_.
+1. In the ``input/omnia_config.yml``, ``input/security_config.yml`` and [optional] ``input/storage_config.yml`` files, provide the `required details <schedulerinputparams.html>`_.
 
 .. note::
     * Use the parameter ``scheduler_type`` in ``input/omnia_config.yml`` to customize what schedulers are installed in the cluster.
     * Without the login node, Slurm jobs can be scheduled only through the manager node.
 
-2. Create an inventory file in the *omnia* folder. Add login node IP address under the manager node IP address under the *[manager]* group, compute node IP addresses under the *[compute]* group, and Login node IP under the *[login]* group,. Check out the `sample inventory for more information <../samplefiles.html>`_.
+2. Create an inventory file in the *omnia* folder. Add login node IP address under the manager node IP address under the *[manager]* group, compute node IP addresses under the *[compute]* group, and Login node IP under the *[login]* group,. Check out the `sample inventory for more information <../../samplefiles.html>`_.
 
 .. note::
      * RedHat nodes that are not configured by Omnia need to have a valid subscription. To set up a subscription, `click here <https://omnia-doc.readthedocs.io/en/latest/Roles/Utils/rhsm_subscription.html>`_.
      * Omnia creates a log file which is available at: ``/var/log/omnia.log``.
      * If only Slurm is being installed on the cluster, docker credentials are not required.
 
-3. To run ``omnia.yml``: ::
+
+3. ``omnia.yml`` is a wrapper playbook comprising of:
+
+    i. ``security.yml``: This playbook sets up centralized authentication (LDAP/FreeIPA) on the cluster. For more information, `click here. <Authentication.html>`_
+    ii. ``scheduler.yml``: This playbook sets up job schedulers (Slurm or Kubernetes) on the cluster.
+    iii. ``storage.yml``: This playbook sets up storage tools like `BeeGFS <BeeGFS.html>`_ and `NFS <NFS.html>`_.
+
+To run ``omnia.yml``: ::
 
         ansible-playbook omnia.yml -i inventory