You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When using SSH launcher on machine A and machine B. if machine a is missing module x and machine b has it: I do dv.execute('import x') the exception lists [n:apply]: ImportError: No module named x where n is the engine number. n should only encompass engine numbers from machine A. But the import error is raised by all engines even though some can import the package.
example:
CompositeError: one or more exceptions from call to method: _execute
[0:apply]: ImportError: No module named tardis
[1:apply]: ImportError: No module named tardis
[2:apply]: ImportError: No module named tardis
[3:apply]: ImportError: No module named tardis
[4:apply]: ImportError: No module named tardis
[5:apply]: ImportError: No module named tardis
[6:apply]: ImportError: No module named tardis
[7:apply]: ImportError: No module named tardis
[8:apply]: ImportError: No module named tardis
[9:apply]: ImportError: No module named tardis
[10:apply]: ImportError: No module named tardis
[11:apply]: ImportError: No module named tardis
[12:apply]: ImportError: No module named tardis
[13:apply]: ImportError: No module named tardis
although rc[13].execute('import tardis') works fine, whereas rc[1].execute('import tardis') throws an error as it should.
The text was updated successfully, but these errors were encountered:
Interesting. I cannot reproduce this. I created a similar situation (some engines have tardis.py, others do not), and trying to do the parallel import results in errors raised on the correct subset of engines.
Perhaps you can give more details? I have no idea what would be relevant, other than making sure you are using up-to-date master.
You know what. I think I might be mixing 0.12.1 with 0.13dev (0.12 on the server side and 0.13.dev on the client side). Is that bad? Are there significant changes?
On 2012-06-25, at 5:31 PM, Min RK wrote:
I just realized the fact that it says 'apply' means you are definitely not using master. Are you using 0.12?
Reply to this email directly or view it on GitHub: #2039 (comment)
When using SSH launcher on machine A and machine B. if machine a is missing module x and machine b has it: I do dv.execute('import x') the exception lists [n:apply]: ImportError: No module named x where n is the engine number. n should only encompass engine numbers from machine A. But the import error is raised by all engines even though some can import the package.
example:
CompositeError: one or more exceptions from call to method: _execute
[0:apply]: ImportError: No module named tardis
[1:apply]: ImportError: No module named tardis
[2:apply]: ImportError: No module named tardis
[3:apply]: ImportError: No module named tardis
[4:apply]: ImportError: No module named tardis
[5:apply]: ImportError: No module named tardis
[6:apply]: ImportError: No module named tardis
[7:apply]: ImportError: No module named tardis
[8:apply]: ImportError: No module named tardis
[9:apply]: ImportError: No module named tardis
[10:apply]: ImportError: No module named tardis
[11:apply]: ImportError: No module named tardis
[12:apply]: ImportError: No module named tardis
[13:apply]: ImportError: No module named tardis
although rc[13].execute('import tardis') works fine, whereas rc[1].execute('import tardis') throws an error as it should.
The text was updated successfully, but these errors were encountered: