squidient 3.4.4__tar.gz → 3.4.5__tar.gz

{squidient-3.4.4 → squidient-3.4.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: squidient
-Version: 3.4.4
+Version: 3.4.5
 Summary: CI/CD framework for alya
 Project-URL: Homepage, https://alya.gitlab.bsc.es/alya/ci-cd/squidient
 Project-URL: Bug Tracker, https://alya.gitlab.bsc.es/alya/ci-cd/squidient/issues

{squidient-3.4.4 → squidient-3.4.5}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "squidient"
-version = "3.4.4"
+version = "3.4.5"
 authors = [
   { name="Damien Dosimont", email="damien.dosimont@gmail.com" },
 ]

{squidient-3.4.4 → squidient-3.4.5}/src/squidient/data/builds/builds.json

@@ -555,6 +555,11 @@
         "platform": "t3.small",
         "system": "aws-ec2-t3.small-x2"
     },
+    "cmake-aws-ec2-t3.small-x6-gnu-14.x.x-i4": {
+        "from": "cmake-aws-ec2-t3.small-gnu-14.x.x-i4",
+        "platform": "t3.small",
+        "system": "aws-ec2-t3.small-x6"
+    },
     "cmake-grace-gnu-14.2.0-i4": {
         "codeCoverage": false,
         "compiler": "gnu",

{squidient-3.4.4 → squidient-3.4.5}/src/squidient/data/configurations/systems.json

@@ -282,6 +282,63 @@
             }
         }
     },
+    "aws-ec2-t3.small-x6": {
+        "host": "t3.small",
+        "path": "/shared/fedora/squidient",
+        "user": "fedora",
+        "project": "",
+        "wait_cloud_init": "t3.small",
+        "platforms": {
+            "t3.small": {
+                "terraform": {
+                    "enable": true,
+                    "sshKeyName": "t3.small"
+                },
+                "account": "",
+                "architecture": {
+                    "tasks_per_node": 2,
+                    "cpus_per_task": 1,
+                    "type": "slurm",
+                    "sjobexitmod": false
+                },
+                "mpirun": {
+                    "openmpi": {
+                        "options": [
+                            "--bind-to core",
+                            "--map-by core",
+                            "--oversubscribe"
+                        ]
+                    }
+                },
+                "jobThreads": 1,
+                "testMaxTasks": 12,
+                "installation": {
+                    "modulePath": "/etc/modulefiles/",
+                    "applicationPath": "/opt/ALYA",
+                    "pythonPath": "/usr/bin"
+                },
+                "alya-mpio-tools": {
+                    "modules": []
+                },
+                "queues": {
+                    "builds": "aws",
+                    "tests": "aws",
+                    "benchmarks": "aws",
+                    "flex": "aws",
+                    "install": "aws",
+                    "default": "aws"
+                },
+                "partition": {},
+                "qos": {},
+                "reservation": {
+                    "enable": false
+                },
+                "host": "",
+                "timeOffset": "00:10:00",
+                "merge": true
+            }
+        }
+    },
     "aws-ec2-t3.small-x3": {
         "host": "t3.small",
         "path": "/shared/fedora/squidient",

squidient-3.4.5/src/squidient/data/terraform/aws-ec2-t3.small-x6/t3.small/.gitignore

@@ -0,0 +1,2 @@
+.terraform*
+terraform.tfstate*

squidient-3.4.5/src/squidient/data/terraform/aws-ec2-t3.small-x6/t3.small/cloud-init-slurm-fedora.sh

@@ -0,0 +1,321 @@
+#!/bin/bash
+set -eux
+
+# ------------------------------------------------------------
+# Role and topology (filled by Terraform via templatefile)
+# ------------------------------------------------------------
+ROLE="${ROLE}"
+HEAD_HOSTNAME="${HEAD_HOSTNAME}"
+COMPUTE_HOSTNAME="${COMPUTE_HOSTNAME}"
+HEAD_PRIVATE_IP="${HEAD_PRIVATE_IP}"
+COMPUTE_NODE_LIST="${COMPUTE_NODE_LIST}"
+
+# ------------------------------------------------------------
+# Hostname setup
+# ------------------------------------------------------------
+if [ "${ROLE}" = "head" ]; then
+  hostnamectl set-hostname "${HEAD_HOSTNAME}"
+else
+  hostnamectl set-hostname "${COMPUTE_HOSTNAME}"
+fi
+
+HOSTNAME="$(hostname)"
+
+# ------------------------------------------------------------
+# Basic networking
+# ------------------------------------------------------------
+if [ "${ROLE}" = "head" ]; then
+  HEAD_IP="$(hostname -I | awk '{print $1}')"
+  if ! grep -q " ${HEAD_HOSTNAME}\b" /etc/hosts; then
+    echo "$HEAD_IP ${HEAD_HOSTNAME}" >> /etc/hosts
+  fi
+else
+  if [ -n "${HEAD_PRIVATE_IP}" ] && ! grep -q " ${HEAD_HOSTNAME}\b" /etc/hosts; then
+    echo "${HEAD_PRIVATE_IP} ${HEAD_HOSTNAME}" >> /etc/hosts
+  fi
+  COMPUTE_IP="$(hostname -I | awk '{print $1}')"
+  if [ -n "$COMPUTE_IP" ] && ! grep -q " ${COMPUTE_HOSTNAME}\b" /etc/hosts; then
+    echo "$COMPUTE_IP ${COMPUTE_HOSTNAME}" >> /etc/hosts
+  fi
+fi
+
+# ------------------------------------------------------------
+# Install base packages
+# ------------------------------------------------------------
+dnf -y update
+
+dnf -y install \
+  slurm slurm-slurmctld slurm-slurmd \
+  munge munge-libs munge-devel \
+  cmake gcc gcc-c++ gcc-gfortran \
+  git make perl wget \
+  python3 python3-pip \
+  diffutils which \
+  tar gzip bzip2 \
+  zlib-devel openssl-devel \
+  boost-devel \
+  libibverbs rdma-core \
+  numactl numactl-devel \
+  environment-modules \
+  nfs-utils
+
+chmod 755 /
+chmod 755 /var
+chmod 755 /var/lib
+
+# ------------------------------------------------------------
+# NFS server + client handling (durci)
+# ------------------------------------------------------------
+mkdir -p /shared
+
+if [ "${ROLE}" = "head" ]; then
+
+  chmod 777 /shared
+
+  cat >/etc/exports <<EOF
+/shared *(rw,sync,no_subtree_check,no_root_squash,fsid=0)
+EOF
+
+  systemctl enable nfs-idmapd || true
+  systemctl start  nfs-idmapd || true
+
+  exportfs -ra
+
+  systemctl enable --now nfs-server
+
+  # Register all compute node addresses after slurmctld starts
+  cat >/usr/local/sbin/refresh-slurm-nodeaddrs.sh << 'EOS'
+#!/bin/bash
+set -eux
+
+HEAD_HOSTNAME="slurm-head"
+HEAD_IP="$(hostname -I | awk '{print $1}')"
+
+# Wait for slurmctld
+for i in $(seq 1 60); do
+  if systemctl is-active --quiet slurmctld; then break; fi
+  sleep 2
+done
+
+# Ensure head is in /etc/hosts
+if ! grep -q " $HEAD_HOSTNAME\b" /etc/hosts; then
+  echo "$HEAD_IP $HEAD_HOSTNAME" >> /etc/hosts
+fi
+
+# Wait for all declared compute nodes, not only the first one.
+for i in $(seq 1 60); do
+  all_ready=true
+
+  for node in ${COMPUTE_NODE_LIST}; do
+    ipfile="/shared/$node.ip"
+    if [ ! -f "$ipfile" ]; then
+      all_ready=false
+      continue
+    fi
+    ip="$(awk '{print $1}' "$ipfile")"
+    if [ -z "$ip" ]; then
+      all_ready=false
+      continue
+    fi
+
+    if ! grep -q " $node\\b" /etc/hosts; then
+      echo "$ip $node" >> /etc/hosts
+    fi
+
+    if ! scontrol update NodeName="$node" NodeAddr="$ip"; then
+      all_ready=false
+    fi
+  done
+
+  if $all_ready; then exit 0; fi
+  sleep 3
+done
+
+echo "Compute node address registration timed out" >&2
+exit 1
+EOS
+
+  chmod +x /usr/local/sbin/refresh-slurm-nodeaddrs.sh
+
+else
+  # Client NFS durci
+  if [ -n "${HEAD_PRIVATE_IP}" ]; then
+
+    if ! grep -q "/shared " /etc/fstab; then
+      echo "${HEAD_PRIVATE_IP}:/shared /shared nfs vers=3,noac,actimeo=0,timeo=600,retrans=5,_netdev 0 0" >> /etc/fstab
+    fi
+
+    for i in $(seq 1 40); do
+      if mountpoint -q /shared; then break; fi
+
+      mount -t nfs -o vers=3,noac,actimeo=0,timeo=600,retrans=5 "${HEAD_PRIVATE_IP}:/shared" /shared && break || true
+      mount -t nfs -o vers=4 "${HEAD_PRIVATE_IP}:/shared" /shared && break || true
+
+      sleep 3
+    done
+
+    if mountpoint -q /shared; then
+      COMPUTE_IP="$(hostname -I | awk '{print $1}')"
+      echo "$COMPUTE_IP" > "/shared/${COMPUTE_HOSTNAME}.ip" || true
+    fi
+  fi
+fi
+
+# ------------------------------------------------------------
+# Munge setup
+# ------------------------------------------------------------
+mkdir -p /etc/munge /var/lib/munge /var/log/munge
+
+if [ "${ROLE}" = "head" ]; then
+  if [ ! -f /etc/munge/munge.key ]; then
+    if command -v create-munge-key >/dev/null 2>&1; then
+      create-munge-key
+    else
+      dd if=/dev/urandom bs=1 count=1024 of=/etc/munge/munge.key
+    fi
+  fi
+  cp /etc/munge/munge.key /shared/munge.key || true
+else
+  for i in $(seq 1 30); do
+    if [ -f /shared/munge.key ]; then
+      cp /shared/munge.key /etc/munge/munge.key
+      break
+    fi
+    sleep 2
+  done
+
+  if [ ! -f /etc/munge/munge.key ]; then
+    if command -v create-munge-key >/dev/null 2>&1; then
+      create-munge-key
+    else
+      dd if=/dev/urandom bs=1 count=1024 of=/etc/munge/munge.key
+    fi
+  fi
+fi
+
+chown munge:munge /etc/munge/munge.key
+chmod 600 /etc/munge/munge.key
+chown -R munge:munge /etc/munge /var/lib/munge /var/log/munge
+chmod 700 /var/lib/munge /var/log/munge
+
+systemctl enable munge
+systemctl start munge
+
+# ------------------------------------------------------------
+# Slurm dynamic config
+# ------------------------------------------------------------
+if ! getent group slurm >/dev/null; then
+  groupadd -r slurm
+fi
+if ! id -u slurm >/dev/null 2>&1; then
+  useradd -r -g slurm -d /var/lib/slurm -s /sbin/nologin slurm
+fi
+
+mkdir -p /var/spool/slurm /var/spool/slurmd /var/log/slurm /var/lib/slurm /etc/slurm
+chown -R slurm:slurm /var/spool/slurm /var/spool/slurmd /var/log/slurm /var/lib/slurm
+
+LOCAL_CPUS="$(nproc)"
+SLURM_CONF="/etc/slurm/slurm.conf"
+
+# Slurm configuration header
+cat >"$SLURM_CONF" <<EOF
+ClusterName=squidient-aws
+ControlMachine=${HEAD_HOSTNAME}
+
+SlurmUser=slurm
+SlurmctldPort=6817
+SlurmdPort=6818
+StateSaveLocation=/var/spool/slurm
+SlurmdSpoolDir=/var/spool/slurmd
+SlurmctldTimeout=300
+SlurmdTimeout=300
+SchedulerType=sched/backfill
+SelectType=select/cons_tres
+
+EOF
+
+# Lignes NodeName
+echo "NodeName=${HEAD_HOSTNAME} CPUs=$LOCAL_CPUS State=UNKNOWN" >>"$SLURM_CONF"
+if [ -n "${COMPUTE_NODE_LIST}" ]; then
+  for node in ${COMPUTE_NODE_LIST}; do
+    echo "NodeName=$node CPUs=$LOCAL_CPUS State=UNKNOWN" >>"$SLURM_CONF"
+  done
+fi
+
+echo "" >>"$SLURM_CONF"
+
+# Partition debug
+PARTITION_NODES="${HEAD_HOSTNAME}"
+if [ -n "${COMPUTE_NODE_LIST}" ]; then
+  for node in ${COMPUTE_NODE_LIST}; do
+    PARTITION_NODES="$PARTITION_NODES,$node"
+  done
+fi
+
+echo "PartitionName=debug Nodes=$PARTITION_NODES Default=YES MaxTime=INFINITE State=UP" >>"$SLURM_CONF"
+
+chown slurm:slurm "$SLURM_CONF"
+
+# ------------------------------------------------------------
+# OpenMPI + alya-mpio-tools (HEAD ONLY)
+# ------------------------------------------------------------
+if [ "${ROLE}" = "head" ]; then
+  OMPI_VERSION="4.1.6"
+  OMPI_PREFIX="/shared/openmpi"
+
+  cd /tmp
+  wget "https://download.open-mpi.org/release/open-mpi/v4.1/openmpi-$OMPI_VERSION.tar.gz"
+  tar -xvf "openmpi-$OMPI_VERSION.tar.gz"
+  cd "openmpi-$OMPI_VERSION"
+
+  ./configure --prefix="$OMPI_PREFIX" --enable-mpirun-prefix-by-default
+  make -j"$(nproc)"
+  make install
+
+  # Use this OpenMPI installation
+  export PATH="$OMPI_PREFIX/bin:$PATH"
+
+  ALYA_MPIO_SRC="/shared/alya-mpio-tools-src"
+  ALYA_MPIO_INSTALL="/shared/alya-mpio-tools"
+
+  git clone --depth 1 https://alya.gitlab.bsc.es/alya/alya-mpio-tools.git "$ALYA_MPIO_SRC" || true
+
+  cd "$ALYA_MPIO_SRC"
+  mkdir -p build
+  cd build
+
+  cmake .. \
+    -DCMAKE_C_COMPILER=mpicc \
+    -DCMAKE_CXX_COMPILER=mpicxx \
+    -DMPIEXEC_PREFLAGS="--allow-run-as-root;--oversubscribe" \
+    -DUSE_VTK=OFF \
+    -DUSE_SFC=OFF \
+    -DCMAKE_INSTALL_PREFIX="$ALYA_MPIO_INSTALL"
+
+  make -j"$(nproc)"
+  make install
+fi
+
+# ------------------------------------------------------------
+# Shared environment for MPI + alya-mpio-tools (ALL NODES)
+# ------------------------------------------------------------
+cat >/etc/profile.d/shared-hpc-env.sh <<'EOF'
+export PATH=/shared/openmpi/bin:/shared/alya-mpio-tools/bin:$PATH
+export LD_LIBRARY_PATH=/shared/openmpi/lib:$LD_LIBRARY_PATH
+EOF
+
+chmod +x /etc/profile.d/shared-hpc-env.sh
+
+# ------------------------------------------------------------
+# Start Slurm daemons
+# ------------------------------------------------------------
+if [ "${ROLE}" = "head" ]; then
+  systemctl enable slurmctld
+  systemctl start slurmctld
+  /usr/local/sbin/refresh-slurm-nodeaddrs.sh >/var/log/refresh-slurm-nodeaddrs.log 2>&1
+else
+  systemctl disable slurmctld || true
+fi
+
+systemctl enable slurmd
+systemctl start slurmd

squidient-3.4.5/src/squidient/data/terraform/aws-ec2-t3.small-x6/t3.small/t3.small.tf

@@ -0,0 +1,170 @@
+terraform {
+  required_version = ">= 1.6.0"
+
+  required_providers {
+    aws = {
+      source  = "hashicorp/aws"
+      version = "~> 5.0"
+    }
+  }
+}
+
+provider "aws" {
+  region = "eu-west-3"
+}
+
+variable "fedora_ami_id" {
+  description = "Fedora Cloud AMI ID in eu-west-3"
+  type        = string
+}
+
+# List of compute node hostnames.
+# Six EC2 instances in total: one head and five compute nodes.
+variable "compute_nodes" {
+  type    = list(string)
+  default = ["slurm-node1", "slurm-node2", "slurm-node3", "slurm-node4", "slurm-node5"]
+}
+
+# Primary compute node name used in slurm.conf generated on the head
+locals {
+  primary_compute = var.compute_nodes[0]
+}
+
+resource "aws_key_pair" "hpc" {
+  key_name   = "hpc-key-x6"
+  public_key = file("${path.module}/.ssh/t3.small.pub")
+}
+
+data "aws_vpc" "default" {
+  default = true
+}
+
+data "aws_subnets" "default" {
+  filter {
+    name   = "vpc-id"
+    values = [data.aws_vpc.default.id]
+  }
+}
+
+resource "aws_security_group" "hpc_ssh" {
+  name        = "hpc-ssh-x6"
+  description = "SSH + internal comms for HPC cluster"
+  vpc_id      = data.aws_vpc.default.id
+
+  # SSH from everywhere (OK for tests, not prod)
+  ingress {
+    from_port   = 22
+    to_port     = 22
+    protocol    = "tcp"
+    cidr_blocks = ["0.0.0.0/0"]
+  }
+
+  # Head <-> compute comms (Slurm/MPI/NFS): allow all inside the VPC
+  ingress {
+    from_port   = 0
+    to_port     = 0
+    protocol    = "-1"
+    cidr_blocks = [data.aws_vpc.default.cidr_block]
+  }
+
+  # Outbound: allow everything (updates, git, etc.)
+  egress {
+    from_port   = 0
+    to_port     = 0
+    protocol    = "-1"
+    cidr_blocks = ["0.0.0.0/0"]
+  }
+}
+
+# ---------------------- HEAD -------------------------
+resource "aws_instance" "slurm_head" {
+  ami           = var.fedora_ami_id
+  instance_type = "t3.small"
+
+  root_block_device {
+    volume_size           = 30
+    volume_type           = "gp3"
+    delete_on_termination = true
+  }
+
+  subnet_id              = data.aws_subnets.default.ids[0]
+  vpc_security_group_ids = [aws_security_group.hpc_ssh.id]
+  key_name               = aws_key_pair.hpc.key_name
+
+  associate_public_ip_address = true
+
+  # Head gets a single "primary" compute hostname for its slurm.conf
+  user_data = templatefile(
+    "${path.module}/cloud-init-slurm-fedora.sh",
+    {
+      ROLE              = "head"
+      HEAD_HOSTNAME     = "slurm-head"
+      COMPUTE_HOSTNAME  = local.primary_compute
+      HEAD_PRIVATE_IP   = "0.0.0.0"
+      COMPUTE_NODE_LIST = join(" ", var.compute_nodes)
+    }
+  )
+
+  tags = {
+    Name = "hpc-slurm-head-fedora-x6"
+  }
+}
+
+# -------------------- COMPUTES (for_each) ------------------------
+resource "aws_instance" "slurm_compute" {
+  for_each      = toset(var.compute_nodes)
+  ami           = var.fedora_ami_id
+  instance_type = "t3.small"
+
+  root_block_device {
+    volume_size           = 30
+    volume_type           = "gp3"
+    delete_on_termination = true
+  }
+
+  subnet_id              = data.aws_subnets.default.ids[0]
+  vpc_security_group_ids = [aws_security_group.hpc_ssh.id]
+  key_name               = aws_key_pair.hpc.key_name
+
+  associate_public_ip_address = true
+
+  depends_on = [aws_instance.slurm_head]
+
+  user_data = templatefile(
+    "${path.module}/cloud-init-slurm-fedora.sh",
+    {
+      ROLE              = "compute"
+      HEAD_HOSTNAME     = "slurm-head"
+      COMPUTE_HOSTNAME  = each.key
+      HEAD_PRIVATE_IP   = aws_instance.slurm_head.private_ip
+      COMPUTE_NODE_LIST = join(" ", var.compute_nodes)
+    }
+  )
+
+  tags = {
+    Name = "hpc-${each.key}-fedora-x6"
+  }
+}
+
+# -------------------- OUTPUTS ------------------------
+
+output "head_public_ip" {
+  description = "Public IP of the Slurm head node"
+  value       = aws_instance.slurm_head.public_ip
+}
+
+output "compute_public_ips" {
+  description = "Public IPs of compute nodes (by hostname)"
+  value       = { for name, inst in aws_instance.slurm_compute : name => inst.public_ip }
+}
+
+output "ssh_head" {
+  value = "ssh -i ./.ssh/t3.small fedora@${aws_instance.slurm_head.public_ip}"
+}
+
+output "ssh_compute_examples" {
+  value = {
+    for name, inst in aws_instance.slurm_compute :
+    name => "ssh -i ./.ssh/t3.small fedora@${inst.public_ip}"
+  }
+}

squidient-3.4.5/src/squidient/data/terraform/aws-ec2-t3.small-x6/t3.small/terraform.tfvars

@@ -0,0 +1,2 @@
+fedora_ami_id = "ami-000a61fce24b26066"
+compute_nodes = ["slurm-node1", "slurm-node2", "slurm-node3", "slurm-node4", "slurm-node5"]