Re: 'ellipsoid' crash

[Edward d'A. <ed...@do...>]

That really is a strange problem.  I wouldn't run relax as root
though.  We've used safe routines, security wise, in relax but running
as root is probably not a good idea.  This is likely to be an issue
with the computer set up and the non-local mount points.

Regards,

Edward



On Thu, Nov 6, 2008 at 2:56 AM, Tyler Reddy <TR...@da...> wrote:
> Yeah it's a permission problem. The same thing happens with XPLOR-NIH writing
> files to the mount directory. The command:
>
> sudo relax full_analaysis.py
>
> does the trick. I'm pretty sure I could also chmod a+w the parent directory or
> just execute the script as root.
>
> Tyler
>
> Quoting Edward d'Auvergne <ed...@do...>:
>
>> Hmmm, that's quite strange.  Does the 'oblate/init' directory exist
>> while the results file is missing?
>>
>> Regards,
>>
>> Edward
>>
>>
>> On Wed, Nov 5, 2008 at 9:42 PM, Tyler Reddy <TR...@da...> wrote:
>>> I just got the same problem with 'oblate', and that folder was removed.
>>> I wonder
>>> if there's a problem with doing this via the NFS mount. Maybe a permissions
>>> thing with directory creation from the python script run on the CPU
>>> of another
>>> machine writing results to the mount.
>>>
>>> Tyler
>>>
>>>
>>>
>>> Quoting Edward d'Auvergne <ed...@do...>:
>>>
>>>> Tyler,
>>>>
>>>> This is an easy one.  Just delete the ellipsoid directory!
>>>>
>>>> The problem is that the full_analysis.py scipt is not very robust.
>>>> It's checking for the 'ellipsoid/init' directory to determine which
>>>> round of optimisation it is up to.  Maybe it would be an idea to
>>>> modify the script to be more robust to handle the directory existing
>>>> yet the final results file being missing for any round.  I would guess
>>>> that relax was killed during its execution, hence the directory exists
>>>> yet the results are missing.  I can't do this at the moment but if you
>>>> would like play with some python coding to implement it yourself, the
>>>> changes could be incorporated back into relax for the benefit of all.
>>>> This may not be simple though, so just deleting the offending
>>>> directory will do the trick.
>>>>
>>>> Regards,
>>>>
>>>> Edward
>>>>
>>>>
>>>> On Wed, Nov 5, 2008 at 6:59 PM, Tyler Reddy <TR...@da...> wrote:
>>>>> Hi,
>>>>>
>>>>> Using the full analysis script, optimization of the 'ellipsoid'
>>>>> model crashes
>>>>> out early with the following output:
>>>>>
>>>>> relax> results.write(file='results', dir='ellipsoid/init/', force=True,
>>>>> compress_type=1)
>>>>> Opening the file 'ellipsoid/init//results.bz2' for writing.
>>>>> RelaxError: Cannot open the file 'ellipsoid/init//results.bz2'.  No
>>>>> such file or
>>>>> directory.
>>>>>
>>>>> I'm not sure why it would try to open this directory though. Isn't
>>>>> it odd that
>>>>> there are 2 forward slashes?
>>>>>
>>>>> Tyler
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> relax (http://nmr-relax.com)
>>>>>
>>>>> This is the relax-users mailing list
>>>>> rel...@do...
>>>>>
>>>>> To unsubscribe from this list, get a password
>>>>> reminder, or change your subscription options,
>>>>> visit the list information page at
>>>>> https://mail.gna.org/listinfo/relax-users
>>>>>
>>>>
>>>> _______________________________________________
>>>> relax (http://nmr-relax.com)
>>>>
>>>> This is the relax-users mailing list
>>>> rel...@do...
>>>>
>>>> To unsubscribe from this list, get a password
>>>> reminder, or change your subscription options,
>>>> visit the list information page at
>>>> https://mail.gna.org/listinfo/relax-users
>>>>
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> relax (http://nmr-relax.com)
>>>
>>> This is the relax-users mailing list
>>> rel...@do...
>>>
>>> To unsubscribe from this list, get a password
>>> reminder, or change your subscription options,
>>> visit the list information page at
>>> https://mail.gna.org/listinfo/relax-users
>>>
>>
>
>
>
>
> _______________________________________________
> relax (http://nmr-relax.com)
>
> This is the relax-users mailing list
> rel...@do...
>
> To unsubscribe from this list, get a password
> reminder, or change your subscription options,
> visit the list information page at
> https://mail.gna.org/listinfo/relax-users
>