Split clean-up into separate job (NOAA-EMC#1906)

Moves the clean-up that was previously done in the archive jobs into
their own separate job. The clean-up is also streamlined considerably
by using only `COM_TOP` instead of going through every template.
There is also additional streamlining/corrections in the function that
does the actual removing.

Some settings used by both jobs were elevated to `config.base`.
Others only needed for cleanup were moved to the new config for
that job.

Also corrects a small error encountered when attempting to rerun an
ensemble forecast.

Resolves NOAA-EMC#583
Resolves NOAA-EMC#1872

jobs/JGLOBAL_CLEANUP

@@ -0,0 +1,17 @@
+#! /usr/bin/env bash
+
+source "${HOMEgfs}/ush/preamble.sh"
+source "${HOMEgfs}/ush/jjob_header.sh" -e "cleanup" -c "base cleanup"
+
+"${HOMEgfs}/scripts/exglobal_cleanup.sh"
+status=$?
+[[ ${status} -ne 0 ]] && exit "${status}"
+
+##########################################
+# Remove the Temporary working directory
+##########################################
+cd "${DATAROOT}" || (echo "${DATAROOT} does not exist. ABORT!"; exit 1)
+[[ ${KEEPDATA} = "NO" ]] && rm -rf "${DATA}"
+
+exit 0
+

jobs/rocoto/arch.sh

@@ -16,5 +16,4 @@ export jobid="${job}.$$"
 "${HOMEgfs}"/jobs/JGLOBAL_ARCHIVE
 status=$?
 
-
 exit "${status}"

jobs/rocoto/cleanup.sh

@@ -0,0 +1,19 @@
+#! /usr/bin/env bash
+
+source "${HOMEgfs}/ush/preamble.sh"
+
+###############################################################
+# Source FV3GFS workflow modules
+. "${HOMEgfs}"/ush/load_fv3gfs_modules.sh
+status=$?
+[[ ${status} -ne 0 ]] && exit "${status}"
+
+export job="cleanup"
+export jobid="${job}.$$"
+
+###############################################################
+# Execute the JJOB
+"${HOMEgfs}"/jobs/JGLOBAL_CLEANUP
+status=$?
+
+exit "${status}"

parm/config/gfs/config.arch

@@ -12,13 +12,4 @@ export ARCH_GAUSSIAN="YES"
 export ARCH_GAUSSIAN_FHMAX=${FHMAX_GFS}
 export ARCH_GAUSSIAN_FHINC=${FHOUT_GFS}
 
-#--online archive of nemsio files for fit2obs verification
-export FITSARC="YES"
-export FHMAX_FITS=132
-[[ "${FHMAX_FITS}" -gt "${FHMAX_GFS}" ]] && export FHMAX_FITS=${FHMAX_GFS}
-
-#--starting and ending hours of previous cycles to be removed from rotating directory
-export RMOLDSTD=144
-export RMOLDEND=24
-
 echo "END: config.arch"

parm/config/gfs/config.base.emc.dyn

@@ -394,6 +394,9 @@ export ARCH_CYC=00           # Archive data at this cycle for warm_start capabil
 export ARCH_WARMICFREQ=4     # Archive frequency in days for warm_start capability
 export ARCH_FCSTICFREQ=1     # Archive frequency in days for gdas and gfs forecast-only capability
 
-export DELETE_COM_IN_ARCHIVE_JOB="YES"   # NO=retain ROTDIR.  YES default in arch.sh and earc.sh.
+#--online archive of nemsio files for fit2obs verification
+export FITSARC="YES"
+export FHMAX_FITS=132
+[[ "${FHMAX_FITS}" -gt "${FHMAX_GFS}" ]] && export FHMAX_FITS=${FHMAX_GFS}
 
 echo "END: config.base"

parm/config/gfs/config.cleanup

@@ -0,0 +1,25 @@
+#! /usr/bin/env bash
+
+########## config.cleanup ##########
+echo "BEGIN: config.cleanup"
+
+# Get task specific resources
+source "${EXPDIR}/config.resources" cleanup
+
+export CLEANUP_COM="YES"   # NO=retain ROTDIR.  YES default in cleanup.sh
+
+#--starting and ending hours of previous cycles to be removed from rotating directory
+export RMOLDSTD=144
+export RMOLDEND=24
+
+# Specify the list of files to exclude from the first stage of cleanup
+# Because arrays cannot be exported, list is a single string of comma-
+# separated values. This string is split to form an array at runtime.
+case ${RUN} in
+	gdas | gfs)	exclude_string="*prepbufr*, *cnvstat*, *atmanl.nc" ;;
+	enkf*)      exclude_string="*f006.ens*" ;;
+	*)			exclude_string="" ;;
+esac
+export exclude_string
+
+echo "END: config.cleanup"

parm/config/gfs/config.resources

@@ -14,7 +14,7 @@ if [[ $# -ne 1 ]]; then
     echo "atmensanlinit atmensanlrun atmensanlfinal"
     echo "landanl"
     echo "aeroanlinit aeroanlrun aeroanlfinal"
-    echo "anal sfcanl analcalc analdiag fcst post vrfy fit2obs metp arch echgres"
+    echo "anal sfcanl analcalc analdiag fcst post vrfy fit2obs metp arch cleanup echgres"
     echo "eobs ediag eomg eupd ecen esfc efcs epos earc"
     echo "init_chem mom6ic ocnpost"
     echo "waveinit waveprep wavepostsbs wavepostbndpnt wavepostbndpntbll wavepostpnt"
@@ -773,6 +773,13 @@ elif [[ ${step} = "arch" || ${step} = "earc" || ${step} = "getic" ]]; then
       eval "export memory_${step}=50GB"
     fi
 
+elif [[ ${step} == "cleanup" ]]; then
+    export wtime_cleanup="01:00:00"
+    export npe_cleanup=1
+    export npe_node_cleanup=1
+    export nth_cleanup=1
+    export memory_cleanup="4096M"
+
 elif [[ ${step} = "stage_ic" ]]; then
 
     export wtime_stage_ic="00:15:00"

scripts/exgdas_enkf_earc.sh

@@ -133,172 +133,4 @@ if [ "${ENSGRP}" -eq 0 ]; then
         "gsistat.${RUN}.${PDY}${cyc}.ensmean"
 fi
 
-
-if [[ "${DELETE_COM_IN_ARCHIVE_JOB:-YES}" == NO ]] ; then
-    exit 0
-fi
-
-###############################################################
-# ENSGRP 0 also does clean-up
-###############################################################
-if [[ "${ENSGRP}" -eq 0 ]]; then
-    function remove_files() {
-        # TODO: move this to a new location
-        local directory=$1
-        shift
-        if [[ ! -d ${directory} ]]; then
-            echo "No directory ${directory} to remove files from, skiping"
-            return
-        fi
-        local exclude_list=""
-        if (($# > 0)); then
-            exclude_list=$*
-        fi
-        local file_list
-        declare -a file_list
-        # Suppress warnings about chained commands suppressing exit codes
-        # shellcheck disable=SC2312
-        readarray -t file_list < <(find -L "${directory}" -type f)
-        if (( ${#file_list[@]} == 0 )); then return; fi
-        for exclude in ${exclude_list}; do
-            echo "Excluding ${exclude}"
-            declare -a file_list_old=("${file_list[@]}")
-            # Suppress warnings about chained commands suppressing exit codes
-            # shellcheck disable=SC2312
-            readarray file_list < <(printf -- '%s\n' "${file_list_old[@]}" | grep -v "${exclude}")
-            if (( ${#file_list[@]} == 0 )); then return; fi
-        done
-
-        for file in "${file_list[@]}"; do
-            rm -f "${file}"
-        done
-        # Remove directory if empty
-        rmdir "${directory}" || true
-    }
-
-    # Start start and end dates to remove
-    GDATEEND=$(${NDATE} -"${RMOLDEND_ENKF:-24}"  "${PDY}${cyc}")
-    GDATE=$(${NDATE} -"${RMOLDSTD_ENKF:-120}" "${PDY}${cyc}")
-
-    while [ "${GDATE}" -le "${GDATEEND}" ]; do
-
-        gPDY="${GDATE:0:8}"
-        gcyc="${GDATE:8:2}"
-
-        if [[ -d ${COM_TOP} ]]; then
-            rocotolog="${EXPDIR}/logs/${GDATE}.log"
-            if [[ -f "${rocotolog}" ]]; then
-                set +e
-                 # Suppress warnings about chained commands suppressing exit codes
-                # shellcheck disable=SC2312
-                testend=$(tail -n 1 "${rocotolog}" | grep "This cycle is complete: Success")
-                rc=$?
-                set_strict
-                if [ "${rc}" -eq 0 ]; then
-                    case ${CDUMP} in
-                        gdas)   nmem="${NMEM_ENS}";;
-                        gfs)    nmem="${NMEM_ENS_GFS}";;
-                        *)
-                            echo "FATAL ERROR: Unknown CDUMP ${CDUMP} during cleanup"
-                            exit 10
-                            ;;
-                    esac
-
-                    readarray memlist< <(seq --format="mem%03g" 1 "${nmem}")
-                    memlist+=("ensstat")
-
-                    for mem in "${memlist[@]}"; do
-                        # Atmos
-                        exclude_list="f006.ens"
-                        # Suppress warnings about chained commands suppressing exit codes
-                        # shellcheck disable=SC2312
-                        templates=$(compgen -A variable | grep 'COM_ATMOS_.*_TMPL')
-                        for template in ${templates}; do
-                            MEMDIR="${mem}" YMD="${gPDY}" HH="${gcyc}" generate_com "directory:${template}"
-                            remove_files "${directory}" "${exclude_list[@]}"
-                        done
-
-                        # Wave
-                        exclude_list=""
-                        # Suppress warnings about chained commands suppressing exit codes
-                        # shellcheck disable=SC2312
-                        templates=$(compgen -A variable | grep 'COM_WAVE_.*_TMPL')
-                        for template in ${templates}; do
-                            MEMDIR="${mem}" YMD="${gPDY}" HH="${gcyc}" generate_com "directory:${template}"
-                            remove_files "${directory}" "${exclude_list[@]}"
-                        done
-
-                        # Ocean
-                        exclude_list=""
-                        # Suppress warnings about chained commands suppressing exit codes
-                        # shellcheck disable=SC2312
-                        templates=$(compgen -A variable | grep 'COM_OCEAN_.*_TMPL')
-                        for template in ${templates}; do
-                            MEMDIR="${mem}" YMD="${gPDY}" HH="${gcyc}" generate_com "directory:${template}"
-                            remove_files "${directory}" "${exclude_list[@]}"
-                        done
-
-                        # Ice
-                        exclude_list=""
-                        # Suppress warnings about chained commands suppressing exit codes
-                        # shellcheck disable=SC2312
-                        templates=$(compgen -A variable | grep 'COM_ICE_.*_TMPL')
-                        for template in ${templates}; do
-                            MEMDIR="${mem}" YMD="${gPDY}" HH="${gcyc}" generate_com "directory:${template}"
-                            remove_files "${directory}" "${exclude_list[@]}"
-                        done
-
-                        # Aerosols (GOCART)
-                        exclude_list=""
-                        # Suppress warnings about chained commands suppressing exit codes
-                        # shellcheck disable=SC2312
-                        templates=$(compgen -A variable | grep 'COM_CHEM_.*_TMPL')
-                        for template in ${templates}; do
-                            MEMDIR="${mem}" YMD="${gPDY}" HH="${gcyc}" generate_com "directory:${template}"
-                            remove_files "${directory}" "${exclude_list[@]}"
-                        done
-
-                        # Mediator
-                        exclude_list=""
-                        # Suppress warnings about chained commands suppressing exit codes
-                        # shellcheck disable=SC2312
-                        templates=$(compgen -A variable | grep 'COM_MED_.*_TMPL')
-                        for template in ${templates}; do
-                            MEMDIR="${mem}" YMD="${gPDY}" HH="${gcyc}" generate_com "directory:${template}"
-                            remove_files "${directory}" "${exclude_list[@]}"
-                        done
-                    done
-                fi
-            fi
-        fi
-
-        # Remove any empty directories
-        YMD=${gPDY} HH=${gcyc} generate_com target_dir:COM_TOP_TMPL
-        target_dir="${ROTDIR:?}/${RUN}.${gPDY}/${gcyc}/"
-        if [[ -d ${target_dir} ]]; then
-            find "${target_dir}" -empty -type d -delete
-        fi
-
-        # Advance to next cycle
-        GDATE=$(${NDATE} +"${assim_freq}" "${GDATE}")
-    done
-fi
-
-# Remove enkf*.$rPDY for the older of GDATE or RDATE
-GDATE=$(${NDATE} -"${RMOLDSTD_ENKF:-120}" "${PDY}${cyc}")
-fhmax=${FHMAX_GFS}
-RDATE=$(${NDATE} -"${fhmax}" "${PDY}${cyc}")
-if [ "${GDATE}" -lt "${RDATE}" ]; then
-    RDATE=${GDATE}
-fi
-rPDY=$(echo "${RDATE}" | cut -c1-8)
-clist="enkfgdas enkfgfs"
-for ctype in ${clist}; do
-    COMIN="${ROTDIR}/${ctype}.${rPDY}"
-    [[ -d ${COMIN} ]] && rm -rf "${COMIN}"
-done
-
-###############################################################
-
-
 exit 0

scripts/exgdas_enkf_fcst.sh

@@ -122,7 +122,9 @@ for imem in $(seq "${ENSBEG}" "${ENSEND}"); do
 
    skip_mem="NO"
    if [[ -f ${EFCSGRP}.fail ]]; then
+      set +e
       memstat=$(grep "MEMBER ${ENSMEM}" "${EFCSGRP}.fail" | grep -c "PASS")
+      set_strict
       [[ ${memstat} -eq 1 ]] && skip_mem="YES"
    fi