Hi Stephen, Thanks very much again for pointing out this problem! I have changed this so that all versions of phenix starting now should run on your machine with '2>&1' instead of '>&' for redirecting stderr to the output file. Let me know if this doesn't do it! All the best, Tom T On Oct 26, 2010, at 9:54 AM, Stephen Graham wrote:

Hi Tom,

Thanks for the info. I think the command is the '>&' redirect of stdout and stderr on line 5 of the generated script. If my sh memory is correct, the same should be achievable with:

phenix.autobuild /path/to/TEMP0/PARAMS_1.eff write_run_directory_to_file=/path/to/TEMP0/INFO_FILE_1 > /path/to/TEMP0/RUN_FILE_1.log 2>&1

i.e. redirect stdout to the RUN_FILE_1.log, and then redirect stderr to the same place using '2>&1'

I've done a quick test and this seems to give the expected behaviour.

Cheers,

Stephen

On 26 October 2010 13:08, Thomas C. Terwilliger wrote:

...

Hi Stephen, I'm sorry about the problem! I put in the "sh" with a bash system as you guessed, and I must have used a non-generic command. I will have to check this out on an ubuntu system.

No, unfortunately there is no easy way to get rid of the sh completely (this is all there because I was trying to get rid of csh which is not in ubuntu at all). You can however go back to the last release version of phenix (which used csh), then download csh and you should be in business.

On the sequence file, both fasta and pir should work...but try just a plain sequence of letters, no headers, and that might solve the gui part of this.

All the best, Tom T

...

...

Hi there,

Apologies if this has been reported before, but the autobuild job I tried to run last night failed after a short while...

The job was started from the command line using:

phenix.autobuild data=../mydata_free.mtz seq_file=../protein.fasta map_file=../ccp4i/protein_phaser.1_parrot2.mtz input_map_labels='parrot.F_phi.F parrot.F_phi.phi None'

And it got to the point where it was trying to submit the three jobs before it died as follows:

Splitting work into 3 jobs and running with 1 processors using sh background=False in /home/stepheng/Projects/my_protein/autobuild/AutoBuild_run_3_/TEMP0

Starting job 1...Log will be: /home/stepheng/Projects/my_protein/autobuild/AutoBuild_run_3_/TEMP0/RUN_FILE_1.log *************ERROR ENDING *******************

******************************************************************************** Failed to carry out AutoBuild_build_cycle:

child process stderr output: command: 'sh /home/stepheng/Projects/my_protein/autobuild/AutoBuild_run_3_/TEMP0/RUN_FILE_1' /home/stepheng/Projects/my_protein/autobuild/AutoBuild_run_3_/TEMP0/RUN_FILE_1.sh: 5: Syntax error: Bad fd number ********************************************************************************

*************ERROR ENDING *******************

I think the problem might be that the RUN_FILE_1.sh script seems to use a bash command/redirection that is not implemented in plain old 'sh' spec, and thus not supported by dash (the lean version of sh that ships with ubuntu)? If I change the #! line of RUN_FILE_1.sh to call /bin/bash rather than /bin/sh the script seems to run to completion...

Is there any easy way to change the script template to use /bin/bash or remove the offending command? I'm running phenix dev-550 on an x86 ubuntu 10.10 machine.

Thanks,

Stephen

P.S. The reason I'm trying to run autobuild from the command line rather from the GUI is that the GUI seems to crash every time I give autobuild a sequence file in either the FASTA or PIR format. Is there another format I should be using?

-- Dr Stephen Graham 1851 Research Fellow Cambridge Institute for Medical Research Wellcome Trust/MRC Building Addenbrooke's Hospital, Hills Road Cambridge, CB2 0XY, UK Phone: +44 1223 762 638 _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

_______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Dr Stephen Graham 1851 Research Fellow Cambridge Institute for Medical Research Wellcome Trust/MRC Building Addenbrooke's Hospital, Hills Road Cambridge, CB2 0XY, UK Phone: +44 1223 762 638

Thomas C. Terwilliger Mail Stop M888 Los Alamos National Laboratory Los Alamos, NM 87545 Tel: 505-667-0072 email: [email protected] Fax: 505-665-3024 SOLVE web site: http://solve.lanl.gov PHENIX web site: http:www.phenix-online.org ISFI Integrated Center for Structure and Function Innovation web site: http://techcenter.mbi.ucla.edu TB Structural Genomics Consortium web site: http://www.doe-mbi.ucla.edu/TB CBSS Center for Bio-Security Science web site: http://www.lanl.gov/cbss

Hi Nat, By 'crash', I mean the GUI just hangs and no longer refreshes. But...I just tested with a sequence file with no headers (named protein.seq) and this worked fine. If I try and load the PIR formatted sequence I get a sensible error message. But if I try and load a FASTA formatted sequence file (called sequence.fasta) the GUI becomes unresponsive. Will send files offline, Stephen On 26 October 2010 16:03, Nathaniel Echols wrote:

On Tue, Oct 26, 2010 at 3:15 AM, Stephen Graham wrote:

...

P.S.  The reason I'm trying to run autobuild from the command line rather from the GUI is that the GUI seems to crash every time I give autobuild a sequence file in either the FASTA or PIR format.  Is there another format I should be using?

As Tom says, a plain sequence and nothing else will work - but I use FASTA and PIR for testing, and they should be okay.  Could you please send me a sequence file that has this problem?  (And by "crash", do you mean "completely quits", or "raises an error message"?)

thanks, Nat _______________________________________________ phenixbb mailing list [email protected] http://phenix-online.org/mailman/listinfo/phenixbb

-- Dr Stephen Graham 1851 Research Fellow Cambridge Institute for Medical Research Wellcome Trust/MRC Building Addenbrooke's Hospital, Hills Road Cambridge, CB2 0XY, UK Phone: +44 1223 762 638