mirror/hdf5
2
0
Fork 0
mirror of https://github.com/HDFGroup/hdf5.git synced 2025-01-18 15:15:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

[svn-r22618]

Browse Source 
Purpose:
  HDFFV-8003 - ph5diff (parallel h5diff): daily test failure on ember intermittently during non comparable test file comparison
  HDFFV-7755 - parallel h5diff : hanging on koala intermittently during non comparable test file comparison

Description:
  non-comparable test intermittently hung on koala and ember, but not on jam. it didn't occur until -np reaches 4 or bigger. it occurred once out of many repeated attempts of the same test.
  There was a incorrectly (mistakenly?) duplicated code in MPI section which caused such hang in a certain condition. The test used more processes than other tests, which increased chance to trigger more undone processes, and such process could enter the incorrect code section and wait for wrong pair of send. it explains why it occurred intermittently according to machine condition and using a certain feature. 
  Removed incorrect code which blocked correct code.
  
Tested: some manually repeated test performed
 jam (linux32-LE), koala (linux64-LE), ostrich (linuxppc64-BE)
This commit is contained in:
Jonathan Kim 2012-08-01 12:07:46 -05:00
parent 840ad09105
commit 99ef5765f5
3 changed files with 5 additions and 16 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

4
release_docs/RELEASE.txt
View File

@ -705,6 +705,10 @@ Bug Fixes since HDF5-1.8.0 release
Tools Tools
----- -----
- ph5diff: Fixed intermittent hang issue on a certain operation in
parallel mode. It was detected by daily test for comparing
non-comparable objects, but it could have occurred in other
operations depend on machine condition. HDFFV-8003 (JKM 2012/08/01)
- h5diff: Fixed test failure for "make check" due to failure of - h5diff: Fixed test failure for "make check" due to failure of
copying test files when performed in HDF5 source tree. Also applied copying test files when performed in HDF5 source tree. Also applied
to other tools. to other tools.

9
tools/h5diff/testh5diff.sh
View File

@ -836,14 +836,7 @@ TOOLTEST h5diff_221.txt -c non_comparables1.h5 non_comparables2.h5 /g2
# entire file # entire file
# All the comparables should display differences. # All the comparables should display differences.
if test -n "$pmode"; then TOOLTEST h5diff_222.txt -c non_comparables1.h5 non_comparables2.h5
# parallel mode:
# skip due to ph5diff hangs on koala (linux64-LE) and ember intermittently.
# (HDFFV-8003 - TBD)
SKIP -c non_comparables1.h5 non_comparables2.h5
else
TOOLTEST h5diff_222.txt -c non_comparables1.h5 non_comparables2.h5
fi
# non-comparable test for common objects (same name) with different object types # non-comparable test for common objects (same name) with different object types
# (HDFFV-7644) # (HDFFV-7644)

8
tools/lib/h5diff.c
View File

@ -1411,14 +1411,6 @@ hsize_t diff_match(hid_t file1_id, const char *grp1, trav_info_t *info1,
options->not_cmp = options->not_cmp | nFoundbyWorker.not_cmp; options->not_cmp = options->not_cmp | nFoundbyWorker.not_cmp;
busyTasks--; busyTasks--;
} /* end if */ } /* end if */
else if(Status.MPI_TAG == MPI_TAG_TOK_RETURN)
{
MPI_Recv(&nFoundbyWorker, sizeof(nFoundbyWorker), MPI_BYTE, Status.MPI_SOURCE, MPI_TAG_DONE, MPI_COMM_WORLD, &Status);
nfound += nFoundbyWorker.nfound;
options->not_cmp = options->not_cmp | nFoundbyWorker.not_cmp;
busyTasks--;
havePrintToken = 1;
} /* end else-if */
else if(Status.MPI_TAG == MPI_TAG_TOK_REQUEST) else if(Status.MPI_TAG == MPI_TAG_TOK_REQUEST)
{ {
MPI_Recv(NULL, 0, MPI_BYTE, Status.MPI_SOURCE, MPI_TAG_TOK_REQUEST, MPI_COMM_WORLD, &Status); MPI_Recv(NULL, 0, MPI_BYTE, Status.MPI_SOURCE, MPI_TAG_TOK_REQUEST, MPI_COMM_WORLD, &Status);