Merge remote-tracking branch 'origin/develop' into fix/c768_hera

* origin/develop: Stage atmospheric backgrounds and UFS cubed-sphere history files (NOAA-EMC#2792) Check that a PR driver is still running before trying to kill it (NOAA-EMC#2799) Feature/get arch adds an empty archive job to GEFS system (NOAA-EMC#2772) Marine DA updates (NOAA-EMC#2802) Revert MSU FIX_DIRs back to glopara (NOAA-EMC#2811) Bugfix for updating label states in Jenkins (NOAA-EMC#2808) Clean-up temporary rundirs - take 2. (NOAA-EMC#2753) Change land surface for HR4 (NOAA-EMC#2787) Run METplus serially and correct the name of prod tasks (NOAA-EMC#2804) Update Java Agent launching script for Jenkins connections (NOAA-EMC#2762) Fix erroneous cdump addition (NOAA-EMC#2803) Update ocean post-processing triggers (NOAA-EMC#2784) Update the gfs_utils repository hash (NOAA-EMC#2801) Add fixes for metplus jobs when gfs_cyc=2 or 4 (NOAA-EMC#2791) Simplify resource-related variables, remove CDUMP where unneeded (NOAA-EMC#2727) Remove f000 from atmos rocoto tasks for replay cases (NOAA-EMC#2778)
DavidHuber-NOAA · Aug 9, 2024 · 638a588 · 638a588
2 parents 9a024e3 + e2c0f06
commit 638a588
Show file tree

Hide file tree

Showing 170 changed files with 2,438 additions and 2,862 deletions.
diff --git a/.github/pull_request_template.md b/.github/pull_request_template.md
@@ -33,6 +33,17 @@
 # Change characteristics
 - Is this a breaking change (a change in existing functionality)? YES/NO
 - Does this change require a documentation update? YES/NO
+- Does this change require an update to any of the following submodules? YES/NO (If YES, please add a link to any PRs that are pending.)
+  - [ ] EMC verif-global
+  - [ ] GDAS
+  - [ ] GFS-utils
+  - [ ] GSI
+  - [ ] GSI-monitor
+  - [ ] GSI-utils
+  - [ ] UFS-utils
+  - [ ] UFS-weather-model
+  - [ ] wxflow
+
 
 # How has this been tested?
 <!-- Please list any test you conducted, including the machine.

diff --git a/ci/Jenkinsfile b/ci/Jenkinsfile
@@ -282,10 +282,10 @@ pipeline {
             steps {
                 script {
                     sh(script: """
-                        labels=\$(gh pr view ${env.CHANGE_ID} --repo ${repo_url} --json labels --jq '.labels[].name')
+                        labels=\$(${GH} pr view ${env.CHANGE_ID} --repo ${repo_url} --json labels --jq '.labels[].name')
                         for label in \$labels; do
                             if [[ "\$label" == *"${Machine}"* ]]; then
-                                gh pr edit ${env.CHANGE_ID} --repo ${repo_url} --remove-label "\$label"
+                                ${GH} pr edit ${env.CHANGE_ID} --repo ${repo_url} --remove-label "\$label"
                             fi
                         done
                     """, returnStatus: true)

diff --git a/ci/cases/gfsv17/ocnanal.yaml b/ci/cases/gfsv17/ocnanal.yaml
@@ -17,7 +17,7 @@ base:
   ACCOUNT: {{ 'HPC_ACCOUNT' | getenv }}
 
 ocnanal:
-  SOCA_INPUT_FIX_DIR: {{ FIXgfs }}/gdas/soca/1440x1080x75/soca
+  SOCA_INPUT_FIX_DIR: {{ HOMEgfs }}/fix/gdas/soca/1440x1080x75/soca
   SOCA_OBS_LIST: {{ HOMEgfs }}/sorc/gdas.cd/parm/soca/obs/obs_list.yaml
   SOCA_NINNER: 100
 

diff --git a/ci/scripts/check_ci.sh b/ci/scripts/check_ci.sh
@@ -50,14 +50,14 @@ fi
 export GH
 
 rocotostat=$(command -v rocotostat)
-if [[ -z ${rocotostat+x} ]]; then
+if [[ -z ${rocotostat} ]]; then
   echo "rocotostat not found on system"
   exit 1
 else
   echo "rocotostat being used from ${rocotostat}"
 fi
 rocotocheck=$(command -v rocotocheck)
-if [[ -z ${rocotocheck+x} ]]; then
+if [[ -z ${rocotocheck} ]]; then
   echo "rocotocheck not found on system"
   exit 1
 else
@@ -70,7 +70,7 @@ pr_list=""
 if [[ -f "${pr_list_dbfile}" ]]; then
   pr_list=$("${HOMEgfs}/ci/scripts/utils/pr_list_database.py" --dbfile "${pr_list_dbfile}" --list Open Running) || true
 fi
-if [[ -z "${pr_list+x}" ]]; then
+if [[ -z "${pr_list}" ]]; then
   echo "no PRs open and ready to run cases on .. exiting"
   exit 0
 fi
@@ -124,7 +124,7 @@ for pr in ${pr_list}; do
 
   for pslot_dir in "${pr_dir}/RUNTESTS/EXPDIR/"*; do
     pslot=$(basename "${pslot_dir}") || true
-    if [[ -z "${pslot+x}" ]]; then
+    if [[ -z "${pslot}" ]]; then
       echo "No experiments found in ${pslot_dir} .. exiting"
       exit 0
     fi

diff --git a/ci/scripts/driver.sh b/ci/scripts/driver.sh
@@ -77,8 +77,9 @@ pr_list=$(${GH} pr list --repo "${REPO_URL}" --label "CI-${MACHINE_ID^}-Ready" -
 
 for pr in ${pr_list}; do
   pr_dir="${GFS_CI_ROOT}/PR/${pr}"
+  [[ ! -d ${pr_dir} ]] && mkdir -p "${pr_dir}"
   db_list=$("${ROOT_DIR}/ci/scripts/utils/pr_list_database.py" --add_pr "${pr}" --dbfile "${pr_list_dbfile}")
-  output_ci_single="${GFS_CI_ROOT}/PR/${pr}/output_single.log"
+  output_ci_single="${pr_dir}/output_single.log"
   #############################################################
   # Check if a Ready labeled PR has changed back from once set
   # and in that case completely kill the previose driver.sh cron
@@ -107,7 +108,9 @@ for pr in ${pr_list}; do
            echo -e "${pstree_out}" | grep -Pow "(?<=\()[0-9]+(?=\))" | xargs kill
         fi
       else
-        ssh "${driver_HOST}" 'pstree -A -p "${driver_PID}" | grep -Eow "[0-9]+" | xargs kill'
+        # Check if the driver is still running on the head node; if so, kill it and all child processes
+        #shellcheck disable=SC2029
+        ssh "${driver_HOST}" "pstree -A -p \"${driver_PID}\" | grep -Eow \"[0-9]+\" | xargs kill || echo \"Failed to kill process with PID: ${driver_PID}, it may not be valid.\""
       fi
       {
         echo "Driver PID: Requested termination of ${driver_PID} and children on ${driver_HOST}"
@@ -141,7 +144,7 @@ pr_list=""
 if [[ -f "${pr_list_dbfile}" ]]; then
   pr_list=$("${ROOT_DIR}/ci/scripts/utils/pr_list_database.py" --dbfile "${pr_list_dbfile}" --list Open Ready) || true
 fi
-if [[ -z "${pr_list+x}" ]]; then
+if [[ -z "${pr_list}" ]]; then
   echo "no PRs open and ready for checkout/build .. exiting"
   exit 0
 fi
@@ -155,7 +158,7 @@ fi
 for pr in ${pr_list}; do
   # Skip pr's that are currently Building for when overlapping driver scripts are being called from within cron
   pr_building=$("${ROOT_DIR}/ci/scripts/utils/pr_list_database.py" --display "${pr}" --dbfile "${pr_list_dbfile}" | grep Building) || true
-  if [[ -z "${pr_building+x}" ]]; then
+  if [[ -n "${pr_building}" ]]; then
       continue
   fi
   id=$("${GH}" pr view "${pr}" --repo "${REPO_URL}" --json id --jq '.id')

diff --git a/ci/scripts/tests/test_rocotostat.py b/ci/scripts/tests/test_rocotostat.py
@@ -25,17 +25,16 @@
     database_destination = os.path.join(testdata_full_path, 'database.db')
     wget.download(database_url, database_destination)
 
-try:
-    rocotostat = which('rocotostat')
-except CommandNotFoundError:
+rocotostat_cmd = which('rocotostat')
+if not rocotostat_cmd:
     raise CommandNotFoundError("rocotostat not found in PATH")
 
-rocotostat.add_default_arg(['-w', os.path.join(testdata_path, 'workflow.xml'), '-d', os.path.join(testdata_path, 'database.db')])
+rocotostat_cmd.add_default_arg(['-w', os.path.join(testdata_path, 'workflow.xml'), '-d', os.path.join(testdata_path, 'database.db')])
 
 
 def test_rocoto_statcount():
 
-    result = rocoto_statcount(rocotostat)
+    result = rocoto_statcount(rocotostat_cmd)
 
     assert result['SUCCEEDED'] == 20
     assert result['FAIL'] == 0
@@ -47,15 +46,15 @@ def test_rocoto_statcount():
 
 def test_rocoto_summary():
 
-    result = rocotostat_summary(rocotostat)
+    result = rocotostat_summary(rocotostat_cmd)
 
     assert result['CYCLES_TOTAL'] == 1
     assert result['CYCLES_DONE'] == 1
 
 
 def test_rocoto_done():
 
-    result = rocotostat_summary(rocotostat)
+    result = rocotostat_summary(rocotostat_cmd)
 
     assert is_done(result)
 
@@ -79,10 +78,10 @@ def test_rocoto_stalled():
         database_destination = os.path.join(testdata_full_path, 'stalled.db')
         wget.download(database_url, database_destination)
 
-    rocotostat = which('rocotostat')
-    rocotostat.add_default_arg(['-w', xml, '-d', db])
+    rocotostat_cmd = which('rocotostat')
+    rocotostat_cmd.add_default_arg(['-w', xml, '-d', db])
 
-    result = rocoto_statcount(rocotostat)
+    result = rocoto_statcount(rocotostat_cmd)
 
     assert result['SUCCEEDED'] == 11
     assert is_stalled(result)

diff --git a/ci/scripts/utils/launch_java_agent.sh b/ci/scripts/utils/launch_java_agent.sh
@@ -1,8 +1,70 @@
 #!/bin/env bash
+
 set -e
 
+# ==============================================================================
+# Script Name: launch_java_agent.sh
+#
+# Description:
+#        This script automates the process of launching a Jenkins agent
+#        on a specified machine. It ensures that the necessary
+#        prerequisites are met, such as the availability of JAVA_HOME,
+#        the Jenkins agent launch directory, and proper authentication
+#        with GitHub.
+#
+#       It then proceeds to check if the Jenkins node is online and
+#       decides whether to launch the Jenkins agent based on the node's
+#       status. The agent is launched in the background,
+#       and its PID is logged for reference.
+#
+# Prerequisites:
+#         JAVA_HOME must be set to a valid JDK installation.
+#         Jenkins agent launch directory must exist and be specified.
+#         GitHub CLI (gh) must be installed and authenticated for messeging
+#           from the Jenkins controller to GitHub PR via shell commands.
+#           Jenkins agent launch directory must exist and be specified.
+#           TODO: Must use GitHub CLI v2.25.1 (newer versoins have issues)
+#           https://github.com/cli/cli/releases/download/v2.25.1/gh_2.25.1_linux_amd64.tar.gz
+#         Jenkins controller URL and authentication token must be provided.
+#         jenkins-secret-file:
+#           Must be present in the Jenkins agent launch directory.
+#           This file contains the secret key for the Jenkins agent
+#           established by the Jenkins administrator for each Node.
+#         jenkins_token:
+#           Must be present in the Jenkins agent launch directory.
+#           This file contains the user authentication token for the Jenkins controller
+#           to use the Remote API. This token can be generated by the user
+#           on the Jenkins controller.
+#         controller_user:
+#           Must be set to the Jenkins controller username corresponing to the jenkins_token.
+#
+# Usage: ./launch_java_agent.sh [now] [-f]
+#        The optional 'now' argument forces the script to launch the Jenkins
+#        agent without waiting before trying again.
+#        The optional 'force' argument forces the script to launch the Jenkins regarless of the node status.
+#
+# ==============================================================================
+
+force_launch="False"
+skip_wait="False"
+while getopts ":fnh" flag; do
+  case "${flag}" in
+    f) force_launch="True";;
+    n) skip_wait="True";;
+    h) echo "Usage: ./launch_java_agent.sh [now] [force]
+Two mutually exclusive optional arguments:
+    -n (now) causes the script to launch the Jenkins agent without waiting before trying again.
+    -f (force) forces the script to launch the Jenkins regarless of its connection status."
+       exit 0 ;;
+    *) echo "Unknown flag: ${flag}"
+       exit 1;;
+  esac
+done
+
 controller_url="https://jenkins.epic.oarcloud.noaa.gov"
-controller_user="terry.mcguinness"
+controller_user=${controller_user:-"terry.mcguinness"}
+controller_user_auth_token="jenkins_token"
+
 HOMEgfs="$(cd "$(dirname  "${BASH_SOURCE[0]}")/../../.." >/dev/null 2>&1 && pwd )"
 host=$(hostname)
 
@@ -13,12 +75,10 @@ host=$(hostname)
 source "${HOMEgfs}/ush/detect_machine.sh"
 case ${MACHINE_ID} in
   hera | orion | hercules | wcoss2)
-    echo "Launch Jenkins Java Controler on ${MACHINE_ID}"
-    ;;
+    echo "Launch Jenkins Java Controler on ${MACHINE_ID}";;
   *)
     echo "Unsupported platform. Exiting with error."
-    exit 1
-    ;;
+    exit 1;;
 esac
 
 LOG=lanuched_agent-$(date +%Y%m%d%M).log
@@ -43,9 +103,16 @@ echo "JAVA VERSION: "
 ${JAVA} -version
 
 export GH="${HOME}/bin/gh"
-command -v "${GH}"
+[[ -f "${GH}" ]] || echo "gh is not installed in ${HOME}/bin"
 ${GH} --version
 
+check_mark=$(gh auth status -t 2>&1 | grep "Token:" | awk '{print $1}') || true
+if [[ "${check_mark}" != "✓" ]]; then
+  echo "gh not authenticating with emcbot token"
+  exit 1
+fi
+echo "gh authenticating with emcbot TOKEN ok"
+
 if [[ -d "${JENKINS_AGENT_LANUCH_DIR}" ]]; then
   echo "Jenkins Agent Lanuch Directory: ${JENKINS_AGENT_LANUCH_DIR}"
 else
@@ -56,22 +123,62 @@ cd "${JENKINS_AGENT_LANUCH_DIR}"
 
 if ! [[ -f agent.jar ]]; then
   curl -sO "${controller_url}/jnlpJars/agent.jar"
+  echo "Updated agent.jar downloaded"
+fi
+
+if [[ ! -f "${controller_user_auth_token}" ]]; then
+   echo "User Jenkins authetication TOKEN to the controller for using the Remote API does not exist"
+   exit 1
 fi
+JENKINS_TOKEN=$(cat "${controller_user_auth_token}")
+
+cat << EOF > parse.py
+#!/usr/bin/env python3
+import json,sys
+with open(sys.argv[1], 'r') as file:
+    data = json.load(file)
+print(data.get('offline','True'))
+EOF
+chmod u+x parse.py
 
-JENKINS_TOKEN=$(cat jenkins_token)
+check_node_online() {
+    rm -f curl_response
+    curl_response=$(curl --silent -u "${controller_user}:${JENKINS_TOKEN}" "${controller_url}/computer/${MACHINE_ID^}-EMC/api/json?pretty=true") || true
+    if [[ "${curl_response}" == "" ]]; then
+       echo "ERROR: Jenkins controller not reachable. Exiting with error."
+       exit 1
+    fi
+    echo -n "${curl_response}" > curl_response
+    ./parse.py curl_response
+}
+
+lauch_agent () {
+    echo "Launching Jenkins Agent on ${host}"
+    command="nohup ${JAVA} -jar agent.jar -jnlpUrl ${controller_url}/computer/${MACHINE_ID^}-EMC/jenkins-agent.jnlp  -secret @jenkins-secret-file -workDir ${JENKINS_WORK_DIR}"
+    echo -e "Launching Jenkins Agent on ${host} with the command:\n${command}" >& "${LOG}"
+    ${command} >> "${LOG}" 2>&1 &
+    nohup_PID=$!
+    echo "Java agent running on PID: ${nohup_PID}" >> "${LOG}" 2>&1
+}
+
+if [[ "${force_launch}" == "True" ]]; then
+  lauch_agent
+  exit
+fi
 
-# 
-offline=$(curl --silent -u "${controller_user}:${JENKINS_TOKEN}" "${controller_url}/computer/${MACHINE_ID^}-EMC/api/json?pretty=true" | grep '\"offline\"' | awk '{gsub(/,/,"");print $3}') || true
-echo "Jenkins Agent offline setting: ${offline}"
+offline=$(set -e; check_node_online)
 
-if [[ "${offline}" == "true" ]]; then
-  echo "Jenkins Agent is offline. Lanuching Jenkins Agent on ${host}"
-  command="nohup ${JAVA} -jar agent.jar -jnlpUrl ${controller_url}/computer/${MACHINE_ID^}-EMC/jenkins-agent.jnlp  -secret @jenkins-secret-file -workDir ${JENKINS_WORK_DIR}"
-  echo -e "Lanuching Jenkins Agent on ${host} with the command:\n${command}" >& "${LOG}"
-  ${command} >> "${LOG}" 2>&1 &
-  nohup_PID=$!
-  echo "Java agent running on PID: ${nohup_PID}" >> "${LOG}" 2>&1
-  echo "Java agent running on PID: ${nohup_PID}"
+if [[ "${offline}" != "False" ]]; then
+   if [[ "${skip_wait}" != "True" ]]; then
+      echo "Jenkins Agent is offline. Waiting 5 more minutes to check again in the event it is a temp network issue"
+      sleep 300
+      offline=$(set -e; check_node_online)
+   fi
+   if [[ "${offline}" != "False" ]]; then
+     lauch_agent
+    else
+      echo "Jenkins Agent is online (nothing done)"
+    fi
 else
   echo "Jenkins Agent is online (nothing done)"
 fi
diff --git a/docs/doxygen/mainpage.h b/docs/doxygen/mainpage.h
@@ -24,7 +24,7 @@ To setup an experiment, a python script <b>\c setup_expt.py</b> (located in <b>\
     usage: setup_expt.py [-h] --pslot PSLOT
                      [--configdir CONFIGDIR] [--idate IDATE] [--icsdir ICSDIR]
                      [--resdetatmos RESDET] [--resensatmos RESENS] [--comroot COMROOT]
-                     [--expdir EXPDIR] [--nens NENS] [--cdump CDUMP]
+                     [--expdir EXPDIR] [--nens NENS] [--run RUN]
 
     Setup files and directories to start a GFS parallel. Create EXPDIR, copy
     config files Create ROTDIR experiment directory structure, link initial
@@ -52,7 +52,7 @@ To setup an experiment, a python script <b>\c setup_expt.py</b> (located in <b>\
                         (default: None)
         --nens       number of ensemble members
                         (default: 80)
-        --cdump      CDUMP to start the experiment
+        --run        RUN to start the experiment
                         (default: gdas)
 
 The above script creates directories <b>\c EXPDIR</b> and <b>\c ROTDIR</b>. It will make links for initial conditions from a location provided via the <b>\c --icsdir</b> argument for a chosen resolution for the control <b>\c --resdetatmos</b> and the ensemble <b>\c --resensatmos</b>. Experiment name is controlled by the input argument <b>\c --pslot</b>. The script will ask user input in case any of the directories already exist. It will copy experiment configuration files into the <b>\c EXPDIR</b> from <b>\c CONFIGDIR</b>.

diff --git a/docs/source/init.rst b/docs/source/init.rst
@@ -384,7 +384,7 @@ The warm starts and other output from production are at C768 deterministic and C
 What files should you pull for starting a new experiment with warm starts from production?
 ------------------------------------------------------------------------------------------
 
-That depends on what mode you want to run -- forecast-only or cycled. Whichever mode, navigate to the top of your ``ROTDIR`` and pull the entirety of the tarball(s) listed below for your mode. The files within the tarball are already in the ``$CDUMP.$PDY/$CYC/$ATMOS`` folder format expected by the system.
+That depends on what mode you want to run -- forecast-only or cycled. Whichever mode, navigate to the top of your ``ROTDIR`` and pull the entirety of the tarball(s) listed below for your mode. The files within the tarball are already in the ``$RUN.$PDY/$CYC/$ATMOS`` folder format expected by the system.
 
 For forecast-only there are two tarballs to pull
 

diff --git a/docs/source/setup.rst b/docs/source/setup.rst
@@ -145,7 +145,7 @@ The following command examples include variables for reference but users should
 
    cd workflow
    ./setup_expt.py gfs cycled --idate $IDATE --edate $EDATE [--app $APP] [--start $START] [--gfs_cyc $GFS_CYC]
-     [--resdetatmos $RESDETATMOS] [--resdetocean $RESDETOCEAN] [--resensatmos $RESENSATMOS] [--nens $NENS] [--cdump $CDUMP]
+     [--resdetatmos $RESDETATMOS] [--resdetocean $RESDETOCEAN] [--resensatmos $RESENSATMOS] [--nens $NENS] [--run $RUN]
      [--pslot $PSLOT] [--configdir $CONFIGDIR] [--comroot $COMROOT] [--expdir $EXPDIR] [--icsdir $ICSDIR]
 
 where:
@@ -170,7 +170,7 @@ where:
    * ``$RESDETOCEAN`` is the resolution of the ocean component of the deterministic forecast [default: 0.; determined based on atmosphere resolution]
    * ``$RESENSATMOS`` is the resolution of the atmosphere component of the ensemble forecast [default: 192]
    * ``$NENS`` is the number of ensemble members [default: 20]
-   * ``$CDUMP`` is the starting phase [default: gdas]
+   * ``$RUN`` is the starting phase [default: gdas]
    * ``$PSLOT`` is the name of your experiment [default: test]
    * ``$CONFIGDIR`` is the path to the config folder under the copy of the system you're using [default: $TOP_OF_CLONE/parm/config/]
    * ``$COMROOT`` is the path to your experiment output directory. Your ``ROTDIR`` (rotating com directory) will be created using ``COMROOT`` and ``PSLOT``. [default: $HOME]