Kernel dies when trying to execute first line of notebook Tutorial 1: Bases and Domains

584 views
Skip to first unread message

Claudio Pierard

unread,
Feb 28, 2017, 9:26:35 PM2/28/17
to Dedalus Users
Hi,

When I try to run the first line of code in the Tutorial1 notebook:

    from dedalus import public as de

the kernel dies immediately, and it doesn't print and error message. I already updated my ipython version, but the kernel continues dying.

I'm using Jupyter version 1.0.0, and my ipython version is 5.3.0.

I'm in a mac os 10.11.6.

Dedalus works fine when I run the examples in the repository.

Any ideas of what might be the cause?

Thank you,

Claudio





Keaton Burns

unread,
Mar 2, 2017, 9:49:25 AM3/2/17
to dedalu...@googlegroups.com
Hi Claudio,

Can you create and run your own jupyter notebooks?  If so, does importing dedalus fail there, too, or only when you try the example notebooks?

-Keaton
--
You received this message because you are subscribed to the Google Groups "Dedalus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dedalus-user...@googlegroups.com.
To post to this group, send email to dedalu...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dedalus-users/12c83331-5b8d-4abf-9601-0bdd92d987cf%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Claudio Pierard

unread,
Mar 2, 2017, 12:35:36 PM3/2/17
to Dedalus Users
Hi Keaton,

It also fails when I create a new notebook and try to import dedalus. It is strange...

Best,

c

Keaton Burns

unread,
Mar 2, 2017, 12:43:45 PM3/2/17
to dedalu...@googlegroups.com
Can you run things besides dedalus in a notebook?  Matplotlib, numpy, scipy, and “from mpi4py import MPI”?

Claudio Pierard

unread,
Mar 2, 2017, 12:48:26 PM3/2/17
to Dedalus Users
Yes, I can run Matplotlib, numpy and scipy, but when I run ""from mpi4py import MPI"" the kernel dies.

Keaton Burns

unread,
Mar 2, 2017, 1:13:54 PM3/2/17
to dedalu...@googlegroups.com
Does it work from the ipython terminal? If not, but does work from the regular python terminal, perhaps jupyter might be using a different python version than the one you manually used to run the other example scripts?

On Mar 2, 2017, at 12:48 PM, Claudio Pierard <c.pi...@ciencias.unam.mx> wrote:

Yes, I can run Matplotlib, numpy and scipy, but when I run ""from mpi4py import MPI"" the kernel dies.

--
You received this message because you are subscribed to the Google Groups "Dedalus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dedalus-user...@googlegroups.com.
To post to this group, send email to dedalu...@googlegroups.com.

Claudio Pierard

unread,
Mar 2, 2017, 1:44:39 PM3/2/17
to Dedalus Users
When I run it from ipython notebook, it appears the error message from openmpi trying to connect to the local network

In [1]: from dedalus import public as de
[MacBook-Pro-de-Claudio.local:03998] [[10693,0],0] ORTE_ERROR_LOG: Bad parameter in file orted/pmix/pmix_server.c at line 262
[MacBook-Pro-de-Claudio.local:03998] [[10693,0],0] ORTE_ERROR_LOG: Bad parameter in file ess_hnp_module.c at line 666
--------------------------------------------------------------------------
It looks like orte_init failed for some reason; your parallel process is
likely to abort.  There are many reasons that a parallel process can
fail during orte_init; some of which are due to configuration or
environment problems.  This failure appears to be an internal failure;
here's some additional information (which may only be relevant to an
Open MPI developer):

  pmix server init failed
  --> Returned value Bad parameter (-5) instead of ORTE_SUCCESS
--------------------------------------------------------------------------
[MacBook-Pro-de-Claudio.local:03997] [[INVALID],INVALID] ORTE_ERROR_LOG: Unable to start a daemon on the local node in file ess_singleton_module.c at line 575
[MacBook-Pro-de-Claudio.local:03997] [[INVALID],INVALID] ORTE_ERROR_LOG: Unable to start a daemon on the local node in file ess_singleton_module.c at line 165
--------------------------------------------------------------------------
It looks like orte_init failed for some reason; your parallel process is
likely to abort.  There are many reasons that a parallel process can
fail during orte_init; some of which are due to configuration or
environment problems.  This failure appears to be an internal failure;
here's some additional information (which may only be relevant to an
Open MPI developer):

  orte_ess_init failed
  --> Returned value Unable to start a daemon on the local node (-127) instead of ORTE_SUCCESS
--------------------------------------------------------------------------
--------------------------------------------------------------------------
It looks like MPI_INIT failed for some reason; your parallel process is
likely to abort.  There are many reasons that a parallel process can
fail during MPI_INIT; some of which are due to configuration or environment
problems.  This failure appears to be an internal failure; here's some
additional information (which may only be relevant to an Open MPI
developer):

  ompi_mpi_init: ompi_rte_init failed
  --> Returned "Unable to start a daemon on the local node" (-127) instead of "Success" (0)
--------------------------------------------------------------------------
*** An error occurred in MPI_Init_thread
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
***    and potentially your MPI job)
[MacBook-Pro-de-Claudio.local:3997] Local abort before MPI_INIT completed completed successfully, but am not able to aggregate error messages, and not able to guarantee that all other processes were killed!


So I did "export TMPDIR=/tmp" before executing the ipython terminal, and it worked!  So I did the same before launching jupyter notebook and everything worked fine.

Keaton Burns

unread,
Mar 6, 2017, 11:47:55 AM3/6/17
to dedalu...@googlegroups.com
Hi Claudio,


I’d try the fixes suggested in that thread (setting TMPDIR, etc.), and also checking how things differ between the notebook and terminals, etc.

Best,
-Keaton
--
You received this message because you are subscribed to the Google Groups "Dedalus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dedalus-user...@googlegroups.com.
To post to this group, send email to dedalu...@googlegroups.com.

Keaton Burns

unread,
Mar 6, 2017, 11:50:38 AM3/6/17
to dedalu...@googlegroups.com
Hi Claudio,

Whoops, your previous e-mail only partially loaded on my laptop, so I hadn’t seen that you tried the TMPDIR fix and it worked.  So please ignore my last message!

Best,
-Keaton
Reply all
Reply to author
Forward
0 new messages