Menu
▾
▴
Re: [Psn-general] SCM problem - restarting an interrupted job
|
From: <jus...@no...> - 2009-04-17 09:18:05
|
Hi Lars, Jeroen, Thanks for the suggestions. We'll give them a shot and feed back. Best Justin Justin Wilkins Senior Modeler Modeling & Simulation (Pharmacology) CHBS, WSJ-027.6.076 Novartis Pharma AG Lichtstrasse 35 CH-4056 Basel Switzerland Phone: +41 61 324 6549 Fax: +41 61 324 1246 Cell: +41 76 561 0949 Email : jus...@no... "Lars Lindbom" <lli...@Ph...> 17/04/2009 10:44 Please respond to "General Discussion about PsN." <psn...@li...> To "Elassaiss - Schaap, J. (Jeroen)" <jer...@sp...>, "Gen eral Discussion about PsN." <psn...@li...> cc phi...@no... Subject Re: [Psn-general] SCM problem - restarting an interrupted job Jeroen, Justin, I recognize the problem as well but I cannot tell for sure how to fix it. I have one suggestion that often helps; try this first and then we can go on to the next step if needed. First, as Jeroen pointed out, create a full copy of the whole scm folder. Then, in the m1 catalog of the step which failed ( e.g. basedir/scm_dir1/scm_dir1/m1), remove the small file called "done". This is a log file for PsN that tells it that it doesn't have to recreate the model files. If you remove it it will recreate the files, but as you don't change anything in the setup, they will be identical to the previous versions. This often helps, but in this case you may have to find which NM_run directory in which the failing file should have been created. The mapping of model and NM_Run folder is found in the model_NMrun_translation.txt file in the modelfit_dir folder of the scm step in question. Delete the files in the NM_run folder that belongs to CLCAGE2.lst. Restart the run as you already did. Let me know how it goes, Thanks, Lars --- ursprungligt meddelande --- Från: "Elassaiss - Schaap, J. (Jeroen)" <jer...@sp...> Ämne: Re: [Psn-general] SCM problem - restarting an interrupted job Datum: 16 april 2009 Tid: 08.30.22 Hi Justin, We ran into similar problems with PsN 2.3.1 and NM6.2. I concluded that it had to do with a file copy from a rundirectory to the main directory. A workaround that worked for us was to restart the jobs as you did earlier. So it might help you to re-restart the job; you would probably want to make a backup of the complete scm directory first to be on the save side... We also use SGE on linux (not RH itself, but a RH-based distribution). Best regards, Jeroen Jeroen Elassaiss-Schaap, PhD Modeling & Simulation Expert Pharmacokinetics, Pharmacodynamics & Pharmacometrics (P3) Early Clinical Research and Experimental Medicine Schering-Plough Research Institute T: +31 41266 9320 _____ From: jus...@no... [mailto:jus...@no...] Sent: Wednesday, 15 April, 2009 11:53 To: psn...@li... Cc: phi...@no... Subject: [Psn-general] SCM problem - restarting an interrupted job Hi all, Hope you can help. We are faced with an SCM job that was interrupted, and is now having difficulty restarting. The command used to try and restart the job was scm -config_file=myrun.scm -p_value=0.01 -do_not_drop=AGE,BMI,SEX,RAC,STY -nm_version=6e -threads=40 -retries=1 -rerun=0 -dir=scm_dir1 (essentially, adding -rerun=0 to the command). We got the following error message: < intervening lines detailing recovery of initial three forward steps removed - see attachment for full history > Starting 38 NONMEM executions. 38 in parallel. D:1 .. D:5 .. D:9 .. D:13 .. D:17 .. D:21 .. D:25 .. D:29 .. D:33 .. D:37 .. D:38 .. done Adding STY on EPR Taking a step forward Starting 37 NONMEM executions. 37 in parallel. D:1 .. D:5 .. D:9 .. D:13 .. D:17 .. D:21 .. D:25 .. D:29 .. D:33 .. D:37 .. done Fatal Error: Trying to access output object, that have no data on file(/CHBS/home/ms/loweph1/DTE/SCM2/scm_dir1/scm_dir1/scm_dir1/scm_dir1/ m1/CLCAGE2.lst) or in memory at lib/output_subs.pm line 815 The first three rounds of the SCM were re-assimilated successfully, but PsN choked on the fourth (CLCAGE2.lst was missing). What is the best way to recover from this? I presume some deletion of subdirectories is called for, but given that it has taken us a week to get this far, we don't want to make any mistakes that force us to have to start from scratch. We're using PsN 2.2.5rc1 on a Red Hat Linux-based SGE grid. Does anyone have any suggestions? Best Justin Justin Wilkins Senior Modeler Modeling & Simulation (Pharmacology) CHBS, WSJ-027.6.076 Novartis Pharma AG Lichtstrasse 35 CH-4056 Basel Switzerland Phone: +41 61 324 6549 Fax: +41 61 324 1246 Cell: +41 76 561 0949 Email : jus...@no... <mailto:jus...@no...> This message and any attachments are solely for the intended recipient. If you are not the intended recipient, disclosure, copying, use or distribution of the information included in this message is prohibited --- Please immediately and permanently delete. ------------------------------------------------------------------------------ Stay on top of everything new and different, both inside and around Java (TM) technology - register by April 22, and save $200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco. 300 plus technical and hands-on sessions. Register today. Use priority code J9JMT32. http://p.sf.net/sfu/p _______________________________________________ Psn-general mailing list Psn...@li... https://lists.sourceforge.net/lists/listinfo/psn-general |
