[mpich-discuss] Problems with "make check" in romio building MPICH-4.2.3

Mon Dec 23 09:25:20 CST 2024

Exactly that. However, a change that should have fixed it (which set FILENAME to “”) was already submitted to “runtests.in” here: https://urldefense.us/v3/__https://github.com/pmodels/mpich/commit/90b0b0614039295d5bde1a7a8b2d430813448467__;!!G_uCfscf7eWS!eipJ0rey6FHts9ZlTwo7nKtSZGUaDecxD2qTlS3IBQjMvmMymCtTZh1nZDJXWQ_HR-SjQ-9qmyy55tY$ 

However, that change isn’t present in the mpich-4.2.3 download. I don’t quite follow how it got lost. But you can see it isn’t present here:

https://urldefense.us/v3/__https://github.com/pmodels/mpich/blob/v4.2.3/src/mpi/romio/test/runtests.in__;!!G_uCfscf7eWS!eipJ0rey6FHts9ZlTwo7nKtSZGUaDecxD2qTlS3IBQjMvmMymCtTZh1nZDJXWQ_HR-SjQ-9qN6_cPTs$ 

Cheers,
Edric.

From: Zhou, Hui <zhouh at anl.gov>
Sent: 23 December 2024 15:13
To: discuss at mpich.org
Cc: Edric Ellis <eellis at mathworks.com>
Subject: Re: Problems with "make check" in romio building MPICH-4.2.3

I see! It is the * in "rm -f ${FILENAME}*" causing the issue. We should either remove the "*" or rename "test_hintfile".

Sorry for being dense. 🙂

Hui

________________________________
From: Edric Ellis via discuss <discuss at mpich.org<mailto:discuss at mpich.org>>
Sent: Monday, December 23, 2024 1:46 AM
To: discuss at mpich.org<mailto:discuss at mpich.org> <discuss at mpich.org<mailto:discuss at mpich.org>>
Cc: Edric Ellis <eellis at mathworks.com<mailto:eellis at mathworks.com>>
Subject: Re: [mpich-discuss] Problems with "make check" in romio building MPICH-4.2.3

Yes, runtests deletes the file! As mentioned below, the runtests scripts states “FILENAME=test”, and then at various stages, runtests executes “rm -f ${FILENAME}*” Cheers, Edric. From: Zhou, Hui <zhouh@ anl. gov> Sent: 17 December 2024 16: 11
ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside your organization.

ZjQcmQRYFpfptBannerEnd

Yes, runtests deletes the file! As mentioned below, the runtests scripts states “FILENAME=test”, and then at various stages, runtests executes “rm -f ${FILENAME}*”

Cheers,

Edric.

From: Zhou, Hui <zhouh at anl.gov<mailto:zhouh at anl.gov>>
Sent: 17 December 2024 16:11
To: Edric Ellis <eellis at mathworks.com<mailto:eellis at mathworks.com>>; discuss at mpich.org<mailto:discuss at mpich.org>
Subject: Re: Problems with "make check" in romio building MPICH-4.2.3

Hi Edric,

Thanks for the test log.

The test-suite.log is complaining "test_hintfile" not exist. This is not a temporary file created during testing, but part of the static test files. Something is causing the tests not able to open "/mathworks/devel/sandbox/eellis/3p-mpich-4.2.3/3p/derived/glnxa64/mpich/src/mpi/romio/test/test_hintfile". Any clues?

Hui

________________________________

From: Edric Ellis <eellis at mathworks.com<mailto:eellis at mathworks.com>>
Sent: Tuesday, December 17, 2024 9:49 AM
To: Zhou, Hui <zhouh at anl.gov<mailto:zhouh at anl.gov>>; discuss at mpich.org<mailto:discuss at mpich.org> <discuss at mpich.org<mailto:discuss at mpich.org>>
Subject: RE: Problems with "make check" in romio building MPICH-4.2.3

I see this: make[2]: Entering directory '/mathworks/devel/sandbox/eellis/3p-mpich-4. 2. 3/3p/derived/glnxa64/mpich/src/mpi/romio/test' FAIL: runtests ============================================================================ Testsuite summary

ZjQcmQRYFpfptBannerStart

This Message Is From an External Sender

This message came from outside your organization.

ZjQcmQRYFpfptBannerEnd

I see this:

make[2]: Entering directory '/mathworks/devel/sandbox/eellis/3p-mpich-4.2.3/3p/derived/glnxa64/mpich/src/mpi/romio/test'

FAIL: runtests

============================================================================

Testsuite summary for ROMIO 4.2.3

============================================================================

# TOTAL: 1

# PASS:  0

# SKIP:  0

# XFAIL: 0

# FAIL:  1

# XPASS: 0

# ERROR: 0

============================================================================

See test/test-suite.log

Please report to discuss at mpich.org<mailto:discuss at mpich.org>

============================================================================

make[2]: *** [Makefile:1167: test-suite.log] Error 1

and the test-suite.log has this:

**** Testing syshints ****

MPI_File_open(MPI_COMM_WORLD, argv[1], MPI_MODE_RDONLY, info_mine, &fh): File does not exist, error stack:

ADIOI_NFS_OPEN(37): File /mathworks/devel/sandbox/eellis/3p-mpich-4.2.3/3p/derived/glnxa64/mpich/src/mpi/romio/test/test_hintfile does not exist

Abort(1) on node 0 (rank 0 in comm 0): application called MPI_Abort(MPI_COMM_WORLD, 1) - process 0

FAIL runtests (exit status: 1)

The file test_hintfile did exist before running the tests.

Cheers,

Edric.

From: Zhou, Hui <zhouh at anl.gov<mailto:zhouh at anl.gov>>
Sent: 13 December 2024 16:04
To: discuss at mpich.org<mailto:discuss at mpich.org>
Cc: Edric Ellis <eellis at mathworks.com<mailto:eellis at mathworks.com>>
Subject: Re: Problems with "make check" in romio building MPICH-4.2.3

Hi Edric,

Could you show the relevant console log for the failure?

Hui

________________________________

From: Edric Ellis via discuss <discuss at mpich.org<mailto:discuss at mpich.org>>
Sent: Friday, December 13, 2024 2:59 AM
To: discuss at mpich.org<mailto:discuss at mpich.org> <discuss at mpich.org<mailto:discuss at mpich.org>>
Cc: Edric Ellis <eellis at mathworks.com<mailto:eellis at mathworks.com>>
Subject: [mpich-discuss] Problems with "make check" in romio building MPICH-4.2.3

Hi mpich-discuss, "make check" is failing with mpich-4. 2. 3, in the src/mpi/romio/test subdirectory. I think it's because this file: https: //urldefense. us/v3/__https: //github. com/pmodels/mpich/blob/4. 2. x/src/mpi/romio/test/runtests. in__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyCb5NBiY$

ZjQcmQRYFpfptBannerStart

This Message Is From an External Sender

This message came from outside your organization.

ZjQcmQRYFpfptBannerEnd

Hi mpich-discuss,

"make check" is failing with mpich-4.2.3, in the src/mpi/romio/test subdirectory. I think it's because this file:

https://urldefense.us/v3/__https://github.com/pmodels/mpich/blob/4.2.x/src/mpi/romio/test/runtests.in__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyCb5NBiY$<https://urldefense.us/v3/__https://github.com/pmodels/mpich/blob/4.2.x/src/mpi/romio/test/runtests.in__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyCb5NBiY$>

still says "FILENAME=test" on line 25. (The test fails because "FILENAME=test" eventually causes "test_hintfile" to be delete before it's needed by one of the tests). There's a change over here:

https://urldefense.us/v3/__https://github.com/pmodels/mpich/commit/90b0b0614039295d5bde1a7a8b2d430813448467__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyKS3SEEk$<https://urldefense.us/v3/__https://github.com/pmodels/mpich/commit/90b0b0614039295d5bde1a7a8b2d430813448467__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyKS3SEEk$>

to make it say "FILENAME=", which I think would fix things.

I'm not sufficiently familiar with github to fathom how that change didn't make it in to 4.2.3 when this change to the Makefile

https://urldefense.us/v3/__https://github.com/pmodels/mpich/commit/6e1e892a04dcbb9079afe0c554d9d2b8b2e99b58__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyatfEXkc$<https://urldefense.us/v3/__https://github.com/pmodels/mpich/commit/6e1e892a04dcbb9079afe0c554d9d2b8b2e99b58__;!!G_uCfscf7eWS!c86jBfO2_nn7840CXToEK-Dl8-y3_afw3ARA6rXELHbuMyXHtcqIGwBNDcw6oUUN8brlyugyatfEXkc$>

(which effectively makes top-level "make check" enter the romio/test subdirectory) did make it in to 4.2.3...

Cheers,

Edric.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20241223/3262331f/attachment-0001.html>