Question: Local galaxy is not run on new Mac OSX (El capitan)
1
gravatar for jaeil.han
3.1 years ago by
jaeil.han • 10
United States
jaeil.han • 10 wrote:

I tried to run local galaxy but cannot do that upon upgrade of MacOSX to El capitan recently.

Here is what I got once I type >sh run.sh

galaxy • 1.8k views
ADD COMMENT • link • modified 3.0 years ago by Martin Čech ♦♦ 4.9k • written 3.1 years ago by jaeil.han • 10
1
gravatar for Wolfgang Maier
3.1 years ago by
Germany
Wolfgang Maier • 600 wrote:

Yes, we've seen the exact same error.

I guess the reason is that Galaxy tries to use a slightly outdated version of amqp.

amqp 1.4.7 fixes the El Capitan incompatibility and should be used instead.

 

To fix this immediately for your current installation, this should work (though I haven't tried it yet):

in the file /Users/jaeilhan/Downloads/galaxy-dev/eggs/amqp-1.4.6-py2.7.egg/amqp/five.py replace the line:

libSystem = ctypes.CDLL('libSystem.dylib')

with:

libSystem = ctypes.CDLL(find_library('libSystem.dylib'))

 

ADD COMMENT • link modified 3.1 years ago • written 3.1 years ago by Wolfgang Maier • 600

oh, and keep the leading whitespace as it is

ADD REPLY • link written 3.1 years ago by Wolfgang Maier • 600

Thank you for your comment.

I tried to replace the line you mentioned, but it still isn't working.

 

ADD REPLY • link written 3.1 years ago by jaeil.han • 10
1
gravatar for Martin Čech
3.0 years ago by
Martin Čech ♦♦ 4.9k
United States
Martin Čech ♦♦ 4.9k wrote:

We are aware of this and are working on the fix. It is possibly caused by Kombu and amqp having trouble on _some_ el capitans.


http://docs.celeryproject.org/projects/kombu/en/latest/changelog.html#version-3-0-27
"Fixed libSystem import error on some OS X 10.11 (El Capitan) installations."

Thanks for reporting

ADD COMMENT • link modified 3.0 years ago • written 3.0 years ago by Martin Čech ♦♦ 4.9k

The fix for 15.10 is here and should be merged shortly once tests pass:

https://github.com/galaxyproject/galaxy/pull/1282

Older versions can simply modify eggs.ini as in that PR and should be able to use the newer eggs (however, you are strongly encouraged to update to the latest stable version, especially if developing in Galaxy).

ADD REPLY • link written 3.0 years ago by Nate Coraor ♦ 3.2k
0
gravatar for jaeil.han
3.1 years ago by
jaeil.han • 10
United States
jaeil.han • 10 wrote:

I reinstalled the galaxy and got something different. (It seemed to work at the beginning)

ADD COMMENT • link modified 3.0 years ago by Martin Čech ♦♦ 4.9k • written 3.1 years ago by jaeil.han • 10

That is still the same error. By reinstalling Galaxy you triggered a fresh download of all software it depends upon including the old version of amqp.

Now try to make the change I suggested and restart (not reinstall) Galaxy.
 

ADD REPLY • link written 3.1 years ago by Wolfgang Maier • 600

I see! However, I made the change you suggested as well but got the same error.

 

ADD REPLY • link written 3.1 years ago by jaeil.han • 10

Unfortunately, I do not have access to our El Capitan test machine right now to verify this.

Can you try to change the line again and post the error message if it appears again? I would be surprised if it's the same as above.

ADD REPLY • link written 3.1 years ago by Wolfgang Maier • 600

I wonder if I did something wrong. Here is the part of the five.py file including the change I made.

if sys.version_info < (3, 3):

    import platform
    SYSTEM = platform.system()

    if SYSTEM == 'Darwin':
        import ctypes
        from ctypes.util import find_library
        libSystem = ctypes.CDLL(find_library('libSystem.dylib'))
        CoreServices = ctypes.CDLL(find_library('CoreServices'),
                                   use_errno=True)
        mach_absolute_time = libSystem.mach_absolute_time
        mach_absolute_time.restype = ctypes.c_uint64
        absolute_to_nanoseconds = CoreServices.AbsoluteToNanoseconds
        absolute_to_nanoseconds.restype = ctypes.c_uint64
        absolute_to_nanoseconds.argtypes = [ctypes.c_uint64]

        def _monotonic():
            return absolute_to_nanoseconds(mach_absolute_time()) * 1e-9

and This is the error message I got. It seems to be same.

ADD REPLY • link modified 3.0 years ago by Martin Čech ♦♦ 4.9k • written 3.1 years ago by jaeil.han • 10

Ah, this is *different*. The original error is gone, but that revealed a new one in /Users/jaeilhan/Downloads/galaxy-dev/eggs/kombu-3.0.24-py2.7.egg/kombu/five.py .

That file suffers from the exact same problem as the previous one so modify it like you did before. Hope that is it.

ADD REPLY • link written 3.1 years ago by Wolfgang Maier • 600
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 16.09
Traffic: 155 users visited in the last hour