[mpich-discuss] osx sierra troubles mpich 3.2

[John DelSignore]

I'm not sure if this is your problem or not, but we found that the Mac OS X update from 10.12.0 to 10.12.1 broke TotalView. The root cause of the problem was security related. In Mac OS X 10.12.1, when a process calls execve() (as is done when a debugger launches a target application), Darwin changes the Mach task and thread ports, and the debugger is now required to use the new ports after it receives the execve() exception event from the operating system. The latest release of TotalView (TotalView 2016.07) support Sierra and its various updates, and it's a parallel debugger designed to with MPICH, Open MPI, and other MPIs.

See: http://www.roguewave.com/products-services/totalview/whats-new

Hope this helps.

Cheers, John D.


On 12/07/16 12:39, Burlen Loring wrote:
> after upgrading to osx sierra and latest xcode some of my regression tests crash in MPI calls, that's as far as I've got as MPI init seems to fail when run in the debugger. See below. Can any one advise how to run the program in the debugger on osx sierrra?
>
>
> (lldb) target create "./bin/test/test_table_reader_distribute"
> Current executable set to './bin/test/test_table_reader_distribute' (x86_64).
> (lldb) settings set -- target.run-args  "/Users/bloring/TECA/../TECA_data/test_tc_candidates_20.bin" "/Users/bloring/TECA/../TECA_data/test_table_reader_distribute_20.bin" "step" "0" "-1" "1"
> (lldb) r
> Process 75610 launched: './bin/test/test_table_reader_distribute' (x86_64)
> [cli_1]: write_line error; fd=8 buf=:cmd=init pmi_version=1 pmi_subversion=1
> :
> system msg for write_line failure : Bad file descriptor
> [cli_1]: Unable to write to PMI_fd
> [cli_1]: write_line error; fd=8 buf=:cmd=get_appnum
> :
> system msg for write_line failure : Bad file descriptor
> Fatal error in MPI_Init_thread: Other MPI error, error stack:
> MPIR_Init_thread(474):
> MPID_Init(152).......: channel initialization failed
> MPID_Init(426).......: PMI_Get_appnum returned -1
> [cli_1]: write_line error; fd=8 buf=:cmd=abort exitcode=1094159
> :
> system msg for write_line failure : Bad file descriptor
> Process 75610 exited with status = 15 (0x0000000f)
> (lldb)
> _______________________________________________
> discuss mailing list     discuss at mpich.org
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/discuss
>

_______________________________________________
discuss mailing list     discuss at mpich.org
To manage subscription options or unsubscribe:
https://lists.mpich.org/mailman/listinfo/discuss