Re: [galaxy-dev] Any hint on how to use Galaxy within virtualenv?
Hi Dannon, We are using Galaxy distribution from February 8, 2013. We didn't have time to update to the April 1st distribution yet. I think the update is planned for next Monday. Is this issue supposed to be solved in the February 8 distribution as well? Or only in the latest (from April 1st)? I don't know if it helps at all, but this issue is not happening on all of our cluster nodes. We had 4 new nodes installed in the last weeks and I think they might not have the exact same OS version. I've checked for the python version anyway, and on our previous nodes (the ones where this issue is not present), we have python 2.6.5, while on the 4 newest nodes, it is python version 2.6.6. The "markupsafe" module is nowhere to be found in the package directory of python 2.6.5, but it is indeed present in the package directory of python 2.6.6 (hence why the conflict I suppose). Thank you for your help! This is much appreciated! Jean-Francois From: Dannon Baker To: Jean-Francois Payotte Cc: Galaxy Dev Date: 19/04/2013 06:47 PM Subject:Re: [galaxy-dev] Any hint on how to use Galaxy within virtualenv? Jean-Francois, Are you running a current revision of Galaxy? This issue should be resolved in recent galaxy releases. -Dannon On Fri, Apr 19, 2013 at 3:47 PM, Jean-Francois Payotte < jean-francois.payo...@dnalandmarks.ca> wrote: Dear Galaxy mailing list, We ran into the same issue described in this post: http://dev.list.galaxyproject.org/markupsafe-module-conflicts-tc4656214.html The exact error message we are getting, with a few different tools is: /Galaxy/galaxy_dist/lib/galaxy/__init__.py:79: UserWarning: Module markupsafe was already imported from /usr/lib64/python2.6/site-packages/markupsafe/__init__.pyc, but /Galaxy/galaxy_dist/eggs/MarkupSafe-0.12-py2.6-linux-x86_64-ucs4.egg is being added to sys.path self.check_version_conflict() >From the solution given in the above-mentioned post, I guess we would have to run Galaxy inside virtualenv. (Details of the procedure to install virtualenv, are given here: http://wiki.galaxyproject.org/Admin/Config/Performance/ProductionServer) However, I am not sure to understand completely how it works. Can we use our already existing Galaxy instance with virtualenv? And how can we achieve this? For the installation of "virtualenv" part, I think we are ok. It is the next step: how to tell Galaxy to run in this virtualenv, that I am not sure about. Thank you in advance for your help! Jean-Francois ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] directory copies as zero-length file from job_working_directory?
Hi Nikhil, I've encountered this issue as well when using the API. Some scripts were using a 5 second sleep default to allow for copying the file to the output directory, but I found that I needed up to 15 seconds (occasionally that's not even enough). Cheers, Rob Rob Leclerc, PhD P: (US) +1-(917)-873-3037 P: (Shanghai) +86-1-(861)-612-5469 Personal Email: rob.lecl...@aya.yale.edu On Mar 28, 2013, at 10:06 PM, Nikhil Joshi wrote: > Hi all, > > So I have a tool that I am writing a wrapper script for that produces a > directory of images as one of its outputs. This directory gets created just > fine in the job_working_directory, but then when the job finishes and the > contents of the job_working_directory are copied, the copy of the images > directory is a zero-length *file*. I seem to recall having this problem > before and finding a solution to it, but now I can't seem to find it. Any > ideas? > > - Nik. > > -- > Nikhil Joshi > Bioinformatics Analyst/Programmer > UC Davis Bioinformatics Core > http://bioinformatics.ucdavis.edu/ > najoshi -at- ucdavis -dot- edu > 530.752.2698 (w) > ___ > Please keep all replies on the list by using "reply all" > in your mail client. To manage your subscriptions to this > and other Galaxy lists, please use the interface at: > http://lists.bx.psu.edu/ > > To search Galaxy mailing lists use the unified search at: > http://galaxyproject.org/search/mailinglists/ ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/
Re: [galaxy-dev] directory copies as zero-length file from job_working_directory?
So I found the solution we used before... is there a reason this code hasn't been changed? Is there a better way to do what I want to do? Here is the solution we used: Yes, you have to modify lib/galaxy/objectstore/__init__.py: In class DiskObjectStore find the function 'update_from_file' (line ~340) and replace: try: shutil.copy(file_name, self.get_filename(obj, **kwargs)) except IOError, ex: log.critical('Error copying %s to %s: %s' % (file_name, self._get_filename(obj, **kwargs), ex)) by: try: if os.path.isdir(file_name): dir_name = self._get_filename(obj, **kwargs) os.remove(dir_name) shutil.copytree(file_name, dir_name) else: shutil.copy(file_name, self.get_filename(obj, **kwargs)) except IOError, ex: log.critical('Error copying %s to %s: %s' % (file_name, self._get_filename(obj, **kwargs), ex)) - Nik. On Thu, Mar 28, 2013 at 7:06 PM, Nikhil Joshi wrote: > Hi all, > > So I have a tool that I am writing a wrapper script for that produces a > directory of images as one of its outputs. This directory gets created > just fine in the job_working_directory, but then when the job finishes and > the contents of the job_working_directory are copied, the copy of the > images directory is a zero-length *file*. I seem to recall having this > problem before and finding a solution to it, but now I can't seem to find > it. Any ideas? > > - Nik. > > -- > Nikhil Joshi > Bioinformatics Analyst/Programmer > UC Davis Bioinformatics Core > http://bioinformatics.ucdavis.edu/ > najoshi -at- ucdavis -dot- edu > 530.752.2698 (w) > -- Nikhil Joshi Bioinformatics Analyst/Programmer UC Davis Bioinformatics Core http://bioinformatics.ucdavis.edu/ najoshi -at- ucdavis -dot- edu 530.752.2698 (w) ___ Please keep all replies on the list by using "reply all" in your mail client. To manage your subscriptions to this and other Galaxy lists, please use the interface at: http://lists.bx.psu.edu/ To search Galaxy mailing lists use the unified search at: http://galaxyproject.org/search/mailinglists/