"UNRECOVERABLE library error" in makeBC run.

[fadzilmnor]

Hi,
First,
I'm working on makeBC, and I did 2 tests. Executed from terminal, test 1 was done with a UM start file with no problem. Second test was done with another UM start file (higher resolution) and this error came up (taken from the log):

set up the namelist
set up the input files
set up the output files
Problem with MAKEBC program
 LIBDMAPP WARNING: Unable to open kgni version file /sys/class/gemini/kgni0/version errno 2

lib-4205 : UNRECOVERABLE library error 
  The program was unable to request more memory space.
 
 =====================================================
 GCOM Version 4.5
 ARCHER -DSerial
 Using precision : 64bit INTEGERs and 64bit REALs
 Built at Fri Nov 15 16:24:01 GMT 2013
 =====================================================
 
 ******************************************
 App ID:23, Name:Make Boundary Conditions
 ------------------------------------------
  - Data size is 64 bit. Program is 64 bit.
  - Program is serial.
 ******************************************
 ##########################################
 Running MAKEBC Utility to create a
 Boundary Dataset from Model Output
 ##########################################
 
STASH_MSTR: /work/n02/n02/hum/vn8.5/ctldata/STASHmaster/STASHmaster_A                                                                                                                                                                                                       
  
 Namelist DUMPBOUND read in 
 &DUMP2BOUND  N_DUMPS = 1, NHOURS = 6, UM_VERSN = 850, LCAL360 = F, NO_LAMS = 1, SUB_HR_INT = 0, L_ADV_WINDS_ON = F, L_PC2 = F, L_MURK = F, L_MCR_QCF2 = F, L_MCR_QRAIN = F, L_MCR_QGRAUP = F, L_DUST = F, L_SO2 = F, L_SO4_AITKEN = F, L_SO4_ACCU = F, L_SO4_DISS = F, L_DMS = F, L_NH3 = F, L_SOOT = F, L_BIOMASS = F, L_OCFF = F, L_NITRATE = F, TRACERS_ACTIVE = 150*0, TR_UKCA_ACTIVE = 150*0, NDUSTBIN_IN = 6, NDUSTBIN_OUT = 6 /

????????????????????????????????????????????????????????????????????????????????
??????????????????????????????????? WARNING ????????????????????????????????????
? Warning in routine: main_makebc
? Warning Code:    -1
? Warning Message: Please remove n_dumps, nhours, and sub_hr_int from DUMP2BOUND.
? Warning generated from processor:     0
????????????????????????????????????????????????????????????????????????????????

 Regular Grid for LBC Area:  1
  
LBC Area  1 : /work/n02/n02/hum/vn6.1/ctldata/vert/vertlevs_G3
 
LBC_STREAM_A and LBC_UNIT_NO_A
 1 stream no  1 unit_no 140
 
FT_Output for Boundary Files.
 Unit No 140 FT_Output Y
 Unit No 141 FT_Output N
 Unit No 142 FT_Output N
 Unit No 143 FT_Output N
 Unit No 144 FT_Output N
 Unit No 145 FT_Output N
 Unit No 146 FT_Output N
 Unit No 147 FT_Output N
  
c_io      :Info: Allocated top layer at 0x0x1168140  
c_io      :init: byteswap: Machine is little endian 
c_io      :init: byteswap: IO will byteswap on this machine 
c_io      :init: top: Initialisation done 
c_io      :layer  1 @ 0x0x1168140 ID: 1 name:Top 
c_io      :layer  2 @ 0x0x1168180 ID: 2 name:Timing 
c_io      :layer  3 @ 0x0x11681c0 ID: 3 name:ByteSwap 
c_io      :layer  4 @ 0x0x1168200 ID: 7 name:LIBC 
IO: Initialised IO
c_io (140):Open: File=test_12km_2.alabcou1 
c_io (140):Open: File is new  
IO: Open: test_12km_2.alabcou1 on unit 140                              
IO: from environment variable BCFIL01
 
Processing dump no   1
  
c_io ( 30):Open: File=/work/n02/n02/py015531/dumps/20120430_00 
c_io ( 30):Open: File exists (8917393408 bytes) 
IO: Open: /work/n02/n02/py015531/dumps/20120430_00 on unit  30                              
IO: from environment variable FILE031
 this is a Dump
 
 Model Grid/Levels in this dump. 
 row_length   =      1024
 rows         =       769
 n_rows       =       768
 model_levels =        70
 wet_levels   =        70
 tr_levels    =         1
 bl_levels    =        50
 ozone_levels =        70
 p_field      =    787456
 u_field      =    786432
 No advected wind lbcs
 No murk lbcs
 No pc2 lbcs
 No tracer LBCs requested in this run
 No UKCA tracer LBCs requested in this run
LBC_CHK_VI : Vertical Interpolation required for LBC Area  1
  
 Dump No   1 : Calling IN_INTF.
 
 Dump No   1 : Calling GEN_INTF.
Gen_Intf: Timestep         0 : Generate Atmos LBCs for Area  1
/work/n02/n02/hum/vn8.5/cce/utils/makebc: line 303: 40898: Abort
MAKEBC output in: /work/n02/n02/py015531/tmp/tmp.eslogin006.40664/makebc_out.py015531.d15041.t121708.40893

The file executed (from the terminal, not qsub) is here:
/work/n02/n02/py015531/makebc/makebc_MalaysiaMay2012_12km_test.tcl
pointing to these files:
/work/n02/n02/py015531/makebc/makebc_runscript_MalaysiaMay2012_12km_test
/work/n02/n02/py015531/makebc/makebc_MalaysiaMay2012_12km_test.nl

Second,
I also did the run using qsub and this error came up:

./makebc_runscript_MalaysiaMay2012_12km_test: line 50: 32774 Illegal instruction     /home/n02/n02/grenvill/makebc -n $NL -i $F0 -o $BD1

I got this error in both tests.

Any ideas how to solve these?

BG.

[grenville]

Hi

Please login to the post processor (ssh espp1 or espp2 - see the ARCHER website).

In your script change

export EXEC=/home/n02/n02/grenvill/qxmakebc64/bin to
export EXEC=/home/n02/n02/grenvill/qxmakebc64_westmere/bin

and try running this a the command line on espp1 or espp2.

You can then change the script to allow submission to the serial queue if needed.

Grenville

[fadzilmnor]

Hi Grenville,

This works, thanks.

Fadzil.