All,
I get an error while reading a hdf5 mesh file that has multiple regions defined in the file. If I only read a single region, the code sometimes crashes and sometimes runs to completion. But, code always crashes when trying to read multiple regions from the mesh. I’m getting in error on my Mac.
Attached are PFLOTRAN’s files and PETSc configure log. I had PETSc download and install HDF5. Any ideas what could be going wrong?
+++++++++++++++++++++++++++++++++++++++++++++
+ PFLOTRAN hash
+++++++++++++++++++++++++++++++++++++++++++++
Merge: 4e46f2f75 aecd6aafa
Author: Matthew Paul <mat...@sandia.gov>
Date: Fri Apr 8 15:31:18 2022 +0000
Merged in matt/wipp_bc_kpc6 (pull request #596)
Correcting WIPP VG (AKA KRP 1 and 8) Derivative
Approved-by: Michael Nole
+++++++++++++++++++++++++++++++++++++++++++++
+++++++++++++++++++++++++++++++++++++++++++++
+ PFLOTRAN error
+++++++++++++++++++++++++++++++++++++++++++++
~/projects/pflotran/pflotran-04-08-2022/src/pflotran/pflotran -pflotranin ugrid_flow.in -flow_snes_monitor -flow_snes_converged_reason -flow_snes_linesearch_type bt -pc_type lu -flow_ksp_type gmres -ksp_monitor
================================
PFLOTRAN Development Version
================================
Opening group: Domain
pflotran card:: REGRESSION
pflotran card:: EOS
"EOS,WATER,DENSITY,CONSTANT units" set to default value.
"EOS,WATER,VISCOSITY,CONSTANT units" set to default value.
pflotran card:: GRID
pflotran card:: FLUID_PROPERTY
"FLUID_PROPERTY,diffusion_coefficient units" set to default value.
pflotran card:: MATERIAL_PROPERTY
Name :: soil1
pflotran card:: CHARACTERISTIC_CURVES
Name :: sf1
pflotran card:: DEBUG
pflotran card:: OUTPUT
"OUTPUT,FORMAT,HDF5,# FILES" set to default value.
pflotran card:: TIME
pflotran card:: REGION
all
pflotran card:: REGION
South
pflotran card:: REGION
Top
pflotran card:: DATASET
pflotran card:: FLOW_CONDITION
initial
pflotran card:: FLOW_CONDITION
South
pflotran card:: FLOW_CONDITION
seepage
pflotran card:: INITIAL_CONDITION
"Initial Condition name" set to default value.
pflotran card:: STRATA
pflotran card:: END_SUBSURFACE
---------- Mesh Quality ----------
Mean Quality: 1.31E+00
Max Quality : 1.80E+00
Min Quality : 1.12E+00
----------------------------------
Extent of Gridded Domain
X: 1.0000000000000000 - 6.0000000000000000
Y: 1.0000000000000000 - 2.0000000000000000
Z: -2.7500000000000000 - 10.000000000000000
Opening hdf5 file: mesh_terrain_5x1x20.h5
Opening hdf5 file: mesh_terrain_5x1x20.h5
pflotran(22404,0x113de4dc0) malloc: Incorrect checksum for freed object 0x7fe7fd1299f0: probably modified after being freed.
Corrupt value: 0x30000000000000fb
pflotran(22404,0x113de4dc0) malloc: *** set a breakpoint in malloc_error_break to debug
Program received signal SIGABRT: Process abort signal.
Backtrace for this error:
#0 0x114b5f28e
#1 0x114b5e9ea
#2 0x7fff71d1c5fc
[1] 22404 abort ~/projects/pflotran/pflotran-04-08-2022/src/pflotran/pflotran -pflotranin
-Gautam
Gautam,
Did this input deck ever work in the past? I am trying to determine if this is a newly introduced error.
Glenn
From: 'Bisht, Gautam' via pflotran-dev <pflotr...@googlegroups.com>
Sent: Friday, April 8, 2022 10:39 AM
To: pflotran-dev <pflotr...@googlegroups.com>
Subject: [pflotran-dev: 6115] HDF5 error
Check twice before you click! This email originated from outside PNNL.
--
You received this message because you are subscribed to the Google Groups "pflotran-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to
pflotran-dev...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/pflotran-dev/CO6PR09MB8536A4483ABA5A05B7D752EAF7E99%40CO6PR09MB8536.namprd09.prod.outlook.com.
Hi Glenn,
I didn’t try this inputdeck with older versions of PFLOTRAN.
I modified the regression_tests/default/regions/ugrid_sideset.in test to read two regions from the .h5 file and PFLOTRAN ran successfully. So, I might be making a mistake in my inputdeck and I will now look into this a bit more. Here is the change that I made to the test.
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++
diff --git a/regression_tests/default/regions/ugrid.txt b/regression_tests/default/regions/ugrid.txt
index b2e539cd3..d3adb314a 100644
--- a/regression_tests/default/regions/ugrid.txt
+++ b/regression_tests/default/regions/ugrid.txt
@@ -51,11 +51,11 @@ REGION all
END
REGION ConditionA
- FILE reg_A.ss
+ FILE ./ugrid_sideset_mesh.h5
END
REGION ConditionB
- FILE reg_B.ss
+ FILE ./ugrid_sideset_mesh.h5
END
REGION bottom
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++
-Gautam
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/PH0PR09MB78360DC5165EAB7007094BA19AEA9%40PH0PR09MB7836.namprd09.prod.outlook.com.
Gautam,
The referenced “ic_terrain_5x1x20.h5” was not sent. I tried renaming the ic_terrain_5x1x20.bin to that filename, and it cannot be read. If I comment out the dataset, the input deck runs (likely has nothing to do with the issue that you are experiencing.
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/CO6PR09MB8536B9ACCA4E750111A71652F7EA9%40CO6PR09MB8536.namprd09.prod.outlook.com.
Here is the initial condition file.
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/PH0PR09MB7836A63346AC028F9014B4879AEA9%40PH0PR09MB7836.namprd09.prod.outlook.com.
Why does the file have .bin as an extension. Shouldn’t it be .h5?
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/CO6PR09MB853675C12D4B21CF39A3A421F7EA9%40CO6PR09MB8536.namprd09.prod.outlook.com.
I mistakenly sent the binary PETSc vec file. Here is the h5 file.
.
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/PH0PR09MB7836E8804CD1E8BFD91ACB509AEA9%40PH0PR09MB7836.namprd09.prod.outlook.com.