New URL for NEMO forge!   http://forge.nemo-ocean.eu

Since March 2022 along with NEMO 4.2 release, the code development moved to a self-hosted GitLab.
This present forge is now archived and remained online for history.
trusting_func.sh in branches/2015/dev_r5092_CNRS_SETTE/NEMOGCM/TRUST – NEMO

source: branches/2015/dev_r5092_CNRS_SETTE/NEMOGCM/TRUST/trusting_func.sh @ 5521

Last change on this file since 5521 was 5521, checked in by nicolasmartin, 9 years ago

dev_r5092_CNRS_SETTE Error management redefinition

  • Property eol-style set to native
  • Property svn:executable set to *
  • Property svn:keywords set to Author Date Id Rev URL
File size: 8.4 KB
Line 
1#!/bin/bash
2
3
4comments() {
5    state=$1; LAST_TIME_STEP=0
6
7    if [ "$state" == 'E R R O R' ]; then
8   [ -e time.step ] && LAST_TIME_STEP=$( cat time.step | tr -d [:space:] )
9   export ${LAST_TIME_STEP}
10    fi
11
12    [ -e ocean.output ] && line=$( grep -m1 -A5 "$state" ocean.output | tr -d '\n' )
13
14    if [[ -e ${REFE_DIR}/model.log && $( diff -q model.log ${REFE_DIR}/model.log ) ]]; then
15   line=$( diff model.log ${REFE_DIR}/model.log  | tr -d '\n' )
16    fi
17
18    printf "Comments\n$line\n" | tee mesg_11_comments_${CFG_USER}_${CFG_ARCH}.txt
19}
20
21
22diff_inputs() {
23    mesg='Same'
24    for file in inputs_list.txt $( ls namelist_* ) $( ls *.xml ); do
25   diff -q $file ${REFE_DIR}/$file > /dev/null
26   [ $? -ne 0 ] && mesg='Different'
27    done
28
29    printf "Input files\n$mesg\n" | tee mesg_09_inputs_${CFG_USER}_${CFG_ARCH}.txt
30}
31
32diff_results() {
33    for file in ocean.output $( ls *.stat ); do
34   [ ! -e ${REFE_DIR}/$file ] && export ST='FAILED' && get_out 7
35   diff -q $file ${REFE_DIR}/$file
36   [ $? -ne 0 ] && export ST='FAILED'
37    done
38}
39
40diff_restart() {
41    [ ! -e  ${REFE_DIR}/time.step ] && export ST='FAILED' && get_out 9
42
43    export LAST_TIME_STEP=$( cat ${REFE_DIR}/time.step | tr -d [:space:] )
44    echo 'Last time step of standard run: '${LAST_TIME_STEP}
45
46    if [    $( find -name "*${LAST_TIME_STEP}_restart*.nc" -print -quit ) \
47    && $( diff -q     time.step ${REFE_DIR}/time.step              )  ]; then
48   base_name=$( find -name "*${LAST_TIME_STEP}_restart*.nc" -print -quit | awk -F/ '{print $NF}' \
49                | sed "s/^\(.*\)$LAST_TIME_STEP\_restart.*/\1$LAST_TIME_STEP\_/"                  )
50
51   for comp in restart restart_ice restart_trc; do
52       file=${base_name}${comp} && printf "$file.nc: "
53
54       nb_dom=$( find -name "${file}_[0-9]*.nc" | wc -l | awk '{print $1}' )
55       if [ ${nb_dom} -gt 1 ]; then
56      ${NEMO_TRBD}/rebuild_nemo -t $NPROC $file ${nb_dom}  > /dev/null
57      [ $? -eq 0 ] && rm -f ${base_name}_${comp}_[0-9]*.nc > /dev/null
58       fi
59
60       if [ -e ${REFE_DIR}/$file.nc ]; then
61       # UNIX `cmp` not suitable (filename & timestamp in .nc file)
62      nc_diff=$( $CDO diffn $file.nc ${REFE_DIR}/$file.nc 2> /dev/null | tail -1 )
63
64      if [ ! -z "${nc_diff}" ]; then
65          export ST='FAILED'
66          printf "$CDO ${nc_diff}\n"
67      else
68          printf "identical\n"
69      fi
70
71       else
72      export ST='FAILED'
73       fi
74
75   done
76
77    else
78   export ST='FAILED'
79    fi
80
81}
82
83get_cpu_time() {
84    real_cpu_time=$( eval "${JOB_INFO} ${JOB_ID} | ${JOB_TIME}" )
85    printf "Real CPU time\n${real_cpu_time}\n" | tee mesg_10_cputime_${CFG_USER}_${CFG_ARCH}.txt
86}
87
88get_inputs() {
89    if [ ! -z "${FORC_TAR}" ]; then
90   cmd_iol="tar -tvf ${NEMO_FORC}/${FORC_TAR}"; cmd_iof="tar -vxf ${NEMO_FORC}/${FORC_TAR}"
91    else
92   cmd_iol="ls ${NEMO_FORC}/*"                ; cmd_iof="\cp ${NEMO_FORC}/* ."
93    fi
94
95    ${cmd_iol} > inputs_list.txt
96    ${cmd_iof} > /dev/null
97}
98
99get_soft_rel() {
100    for rel in $CDO $COMPILER $MPI $NETCDF; do
101                                     arch_rel=$( echo $LOADEDMODULES | sed  "s#.*$rel/\([^:]*\).*#\1#" )
102   [ -z "${arch_rel}"  ]     && arch_rel=$( echo $PATH          | sed  "s#.*$rel/\([^/]*\).*#\1#" )
103   [ "$rel" == "$COMPILER" ] && arch_rel=$( $rel --version      | grep -m1 -o ' [0-9.]* '         )
104   echo $rel ${arch_rel} >> model.log
105    done
106
107    sed -n 4p model.log | sed 'i\Fortran compiler' > mesg_06_compiler_${CFG_USER}_${CFG_ARCH}.txt
108    sed -n 5p model.log | sed 'i\MPI libs'         > mesg_07_mpi_${CFG_USER}_${CFG_ARCH}.txt
109    sed -n 6p model.log | sed 'i\NetCDF libs'      > mesg_08_netcdf_${CFG_USER}_${CFG_ARCH}.txt
110}
111
112get_nemo_rev() {
113    svn_opt='status'
114
115    # If -v|--version option has been set, modify svn command
116    if [[    $( echo ${NEMO_VERS} | grep  "HEAD\|up\|update"   ) \
117     || $( echo ${NEMO_VERS} | tr -d '[:alpha:][:punct:]' )  ]]; then
118    rev=$( echo ${NEMO_VERS} | tr -d '[:alpha:][:punct:]' ) || rev='HEAD'
119   svn_cmd='svn update -r '$rev 
120    else
121   svn_cmd='svn '${svn_opt}
122    fi
123
124    rev=0
125    for dir in ${NEMO_ARCH} ${NEMO_CONF} ${NEMO_ENGI} \
126               ${NEMO_EIOI} ${NEMO_EFCM}              \
127          ${NEMO_TCMP} ${NEMO_TRBD}              \
128          ${DIR_XIOS}                             ; do
129
130   # For time being, just get revision number from XIOS
131   if [ "$dir" == "${DIR_XIOS}" ]; then
132       rev_loc=$( svn info $dir | awk '(NR == 5) {print $NF}' )
133       echo   'XIOS '${rev_loc} >> model.log
134       printf "XIOS rev.\n"     >  mesg_05_xios_${CFG_USER}_${CFG_ARCH}.txt
135       printf "<a href=\"https://forge.ipsl.jussieu.fr/ioserver/changeset/${rev_loc}\" target=\"_blank\">${rev_loc}</a>" \
136                                >> mesg_05_xios_${CFG_USER}_${CFG_ARCH}.txt
137       continue
138   fi
139
140   echo ${svn_cmd} $dir': '$( ${svn_cmd} $dir )
141   rev_loc=$( svn info $dir | awk '(NR == 5) {print $NF}' )
142   (( ${rev_loc} >= $rev )) && rev=${rev_loc}
143    done
144
145    echo   'NEMOGCM '$rev   >> model.log
146    printf "NEMOGCM rev.\n" >  mesg_04_nemogcm_${CFG_USER}_${CFG_ARCH}.txt
147    printf "<a href=\"https://forge.ipsl.jussieu.fr/nemo/changeset/$rev\" target=\"_blank\">$rev</a>" \
148                       >> mesg_04_nemogcm_${CFG_USER}_${CFG_ARCH}.txt
149}
150
151mesg_make() {
152    paste -d ';' mesg_*_${CFG_USER}_${CFG_ARCH}.txt > trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.txt
153    cat                                               trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.txt
154}
155
156mesg_publish() {
157 
158if [ $PUBLISH -eq 1 ]; then
159
160  if [ -f ${REFE_DIR}/trusting_${CFG_USER}_${CFG_ARCH}.txt ]; then cmd='tail -1'; else cmd='cat'; fi
161
162  $cmd trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.txt >> ${REFE_DIR}/trusting_${CFG_USER}_${CFG_ARCH}.txt
163
164  # Send mail only when FAILED
165  if [[ ! -z $EMAIL && "$ST" == 'FAILED' ]]; then
166      cat << END_MAIL > trusting.mail
167XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
168
169Dear $USER,
170
171The trusting sequence for ${REFE_CONF} has failed.
172Directory: ${TEST_DIR}
173Archive: ${REFE_DIR}/trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.tar.gz
174
175XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
176
177END_MAIL
178
179      if [ -e ${TEST_DIR}/trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.txt ]; then
180     cat ${TEST_DIR}/trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.txt  >> trusting.mail
181      fi
182      mail -s "[trusting ${REFE_CONF}] $ST $ERR" $EMAIL  <  trusting.mail
183  fi
184
185fi
186}
187
188print_step() { printf "Step.....\n$1\n"; }
189
190get_out() {
191    ERR=$1
192
193    printf "Status\n$ST\n"  > mesg_02_status_${CFG_USER}_${CFG_ARCH}.txt
194
195    if [ "$ST" == 'FAILED' ]; then
196   cd ${TEST_DIR}
197
198   if [ ! -e mesg_09_inputs_${CFG_USER}_${CFG_ARCH}.txt ]; then
199       printf "Input files\n\n"   > mesg_09_inputs_${CFG_USER}_${CFG_ARCH}.txt
200   fi
201
202   if [ ! -e mesg_10_cputime_${CFG_USER}_${CFG_ARCH}.txt ]; then
203       printf "Real CPU time\n\n" > mesg_10_cputime_${CFG_USER}_${CFG_ARCH}.txt
204   fi
205
206   if [ ! -e mesg_11_comments_${CFG_USER}_${CFG_ARCH}.txt ]; then
207       printf "Comments\n\n"      > mesg_11_comments_${CFG_USER}_${CFG_ARCH}.txt
208   fi
209
210        # Error identification
211   case $ERR in
212       # Compilation
213       1) ERR='XIOS compilation failed'            ;;  2) ERR='NEMO compilation failed'            ;;
214       # Submission
215       3) ERR='Missing input files'                 ;;  4) ERR='Job submission error'               ;;
216       # Running
217       5) ERR='nemo.exe crashed at '${LAST_TIME_ERR};;  6) ERR='Exceeded time limit '${TIME_LIMI}'h';;
218       # Results
219       7) ERR='Missing previous outputs '           ;;  8) ERR='New outputs  differ/missing'        ;;
220       9) ERR='Missing previous restarts'           ;; 10) ERR='New restarts differ/missing'        ;;
221       # Undefined
222       *) ERR='Unknown error'                       ;;
223   esac
224
225   export ERR
226    fi
227
228    printf "Result.....\n$ERR\n" > mesg_03_result_${CFG_USER}_${CFG_ARCH}.txt
229
230    # Save tested configuration if trusting failed
231    [ "$ST" == 'FAILED' ] && tar -czf ${REFE_DIR}/trusting_${DATESTR}_${CFG_USER}_${CFG_ARCH}.tar.gz *
232
233    mesg_make
234    mesg_publish
235
236    exit 1
237}
238
239job_pending() {
240    time_elapsed=0; time_increment=30
241
242    sleep ${time_increment}
243    while [[ $( ${JOB_LIST} | grep ${JOB_ID} ) && ${time_elapsed} -lt ${TIME_LIMI} ]]; do
244   printf "\n####################################################" >> computation.log
245   ${JOB_INFO} ${JOB_ID}                                           >> computation.log
246   sleep ${time_increment}
247   time_elapsed=$(( ${time_elapsed} + ${time_increment} ))
248    done
249    sleep ${time_increment}
250
251    if [ ${time_elapsed} -eq ${TIME_LIMI} ]; then
252   ${JOB_DELE} ${JOB_ID} &> /dev/null
253   TIME_LIMIT=$(( ${TIME_LIMI} / 3600 ))
254   export ${TIME_LIMIT}
255   get_out $ST ERR=6
256    fi
257}
Note: See TracBrowser for help on using the repository browser.