Thanks for your quick response, my colleagues of a small group are waiting for using galaxy, so it is better if we can restart it asap.
1. If I try to set up a virtualenv, do I need to recompile "bx-python"? I am not familiar with "Python", if we need to do it, could you please let me know how to do it step by step?
2. I checked your instruction you mentioned, will I follow the following steps (a, b, and c)? If it is, I will try immediately.
nate@weyerbacher% wget http://bitbucket.org/ianb/virtualenv/raw/tip/virtualenv.py
Resolving bitbucket.org... 22.214.171.124, 126.96.36.199
Connecting to bitbucket.org|188.8.131.52|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 68,601 (67K) [text/x-python]
100%[====================================>] 68,601 --.--K/s
11:18:05 (729.46 KB/s) - `virtualenv.py' saved [68601/68601]
(b) run command as follows
nate@weyerbacher% /usr/bin/python2.6 virtualenv.py --no-site-packages galaxy_env
New python executable in galaxy_env/bin/python2.6
Also creating executable in galaxy_env/bin/python
(c)run the following 3 commands
nate@weyerbacher% . ./galaxy_env/bin/activate
nate@weyerbacher% cd galaxy_dist
nate@weyerbacher% sh run.sh
Yan Luo wrote:
> Dear Nate,
> Why did you say we have 3 paths? We have a link fromOkay, this should be fine.
> lrwxrwxrwx 1 kangtu admin 36 2010-11-01 17:07 galaxy-dist ->
> from "/home/kangtu/data", we have other link as follows.
> lrwxrwxrwx 1 kangtu kangtu 28 2010-10-04 10:25 data ->
> They worked well before, when we just restarted, it has such problem, I
> couldn't start. By the way,
> 1. I have found some files as "root:root" under
> "/home/kangtu/Downloads/galaxy-dist" including its subfolders' files, so I
> changed all them as "kangtu/admin", there is only one user right now (we
> use postgresql db).
You would need to recompile whatever depends on numpy, which in this
> 2. My colleague told me that he upgraded "numpy" to new version recently, is
> it the problem? how can we fix it? I checked online, someone said we need to
> recompile something (http://projects.scipy.org/numpy/ticket/1320).
case is bx-python. However, we provide our own version of numpy 1.3.
The reason that version is not being used is due to a bug in our
dependency management code which was just fixed.
There are two ways to solve this problem. The first would be to set up
a virtualenv as suggested in my last email.
The second would be to pull our in-development code, but be aware that
this has undergone very little testing at this point. You can pull the
most recent changes with:
% hg pull -u http://bitbucket.org/galaxy/galaxy-central/
> Looking forward to hearing from you.
> On Thu, Feb 10, 2011 at 10:26 PM, Nate Coraor <firstname.lastname@example.org> wrote:
> > Yan Luo wrote:
> > > Dear Nate,
> > >
> > > Please find the following errors, it is the newest, and our restart is
> > > stopped. I hope we can get the solution from you as soon as possible.
> > Hi Yan,
> > Your last 4 messages included 3 different paths to Galaxy. It seems as
> > though there are quite a few copies of Galaxy in use here.
> > To ensure there are no permissions problems, please do the following,
> > all as the same non-root user:
> > 1. Remove the ~/.python-eggs directory and all of its contents.
> > 2. Clone a new copy of Galaxy.
> > 3. Start Galaxy using `./run.sh`
> > And report back any errors which persist. If the numpy error below
> > remains, I suggest setting up a Python virtualenv as described on:
> > http://usegalaxy.org/production
> > --nate