locked
Access Denied Error...[fixed] But, closes unexpectadly now. RRS feed

  • Question

  • User252006738 posted

    Alright, followed the instructions, gave ruby the appropriate permissions, setup them all the way I believe they are supposed to be (most was pre-done by plesk.  Just had to apply all the way down the app tree).  Still not getting it to work.

     Any ideas?
     

    Thursday, March 1, 2007 6:55 PM

All replies

  • User989702501 posted

    Are you referring to this ?
    http://mvolo.com/blogs/serverside/archive/2007/02/18/10-steps-to-get-Ruby-on-Rails-running-on-Windows-with-IIS-FastCGI.aspx

    if yes, try post your question at Mike's blog.

    Friday, March 2, 2007 12:56 AM
  • User252006738 posted

    I have the Acces Denied error fixed.  The Plesk CP utils resticted it's own access to the ruby folder.  Fixed by giving it read and execute permissions.  Now however, I get a different error with no logging that I can find to describe what is going on.

    Error 0x80004005 occurred processing request.

    The FastCGI process exited unexpectedly.

    Any thoughts?
     

    Wednesday, March 7, 2007 10:58 AM
  • User-679828332 posted
    This means that the ruby interpreter exited in the middle of the request. You could try collecting a process monitor (www.sysinternals.com) log of the request.
    Wednesday, March 7, 2007 11:46 AM
  • User252006738 posted
    According to Process Monitor, it loads and exits successfully.  Any other thoughts?
    Thursday, March 8, 2007 2:17 AM
  • User-679828332 posted

    Are there any file system or registry "access denieds" for ruby.exe?

    If not, then next thing I'd try is adding logging to your ruby app: create a log file on the disk and see how far you get..

    Thursday, March 8, 2007 12:42 PM
  • User252006738 posted

    Well, after configuring it again.  Now I get an "Invalid Index" from fastcgi.  Ruby never gets loaded this time.  I feel like I took a step back this time.

     When Ruby was loading, there were some access denieds.
     

    Friday, March 9, 2007 8:14 PM
  • User-679828332 posted

    Hi rjamesh,

    Invalid index means one part of your config references another part that does not exist. If you share your fcgiext.ini config file I can tell you what's wrong.

    Cheerz,

    Rick.

    Saturday, March 10, 2007 12:57 PM
  • User252006738 posted

    Everything looks alright though.

     [Types]
    *:26228=DHM

    [DHM]
    ExePath=c:\ruby\bin\ruby.exe
    Arguments=[path removed due to contract restrictions.  But is correct in file.]\test\public\dispatch.fcgi
    IgnoreDirectories=0
    IgnoreExistingFiles=1
    QueueLength=1000
    MaxInstances=4
    InstanceTimeout=30
    InstanceMaxRequests=200 

    Monday, March 12, 2007 12:01 AM
  • User-119923882 posted

    There are two possible ways that this config could result in ERROR_INVALID_INDEX:

    1) If the site ID is incorrect for the request, no match will be found in the [types] section, and this error will result.

    2) If you are running IIS 5.1 and earlier, wildcard script maps are not supported.  Based on the implementation of the code, the above [types] section would have no valid mappings are would also result in this error.

    I hope that this helps,

    -Wade

    Monday, March 12, 2007 5:06 PM
  • User252006738 posted
    Well.  The Site ID is correct (tripple checked) and the server is IIS6.0 on 2k3.  Any other ideas?  If it helps any, using Plesk to do site admin.  It's caused me enough other problems that it might be part of it.
    Friday, March 16, 2007 10:51 AM
  • User-119923882 posted

    The code that generates the ERROR_INVALID_INDEX error can really only be reached in a couple of ways, and they both have to do with not being able to match the request to a line in [types] or not being able to match the line in [types] with another section in the ini file.  I don't see how your application could cause it, since the error is happening before we even attempt to load it.

    As for other ideas, it is possible that you're editing the wrong copy of fcgiext.ini.  We ran into there here a few days ago when a test machine ended up with two different copies of fcgiext.dll and was editing the ini file for the one that wasn't script mapped.

    Another way that this could happen is if you are on a 64 bit OS and running a 32 bit worker process, make sure that you are not getting fooled by the file system mapping between system32 and syswow64.  To check this, look under both \windows\system32 and \windows\syswow64 to see if there are fcgiext.dll files under both of them (be sure to check child directories, too.)  If there are, try editing both fcgiext.ini files.

    Finally, if you get desperate, try attaching the ntsd.exe debugger to the worker process.  When you attach it to w3wp.exe, it should list the loaded dlls, and it will probably list a full path to the fcgiext.dll that's actually getting loaded.  That will be the same path to the fcgiext.ini file that is being used.

    Please let me know how it goes.

    Thanks,

    -Wade

    Friday, March 16, 2007 4:49 PM
  • User252006738 posted
    Well, I started fresh this time.  No longer get the invalid index, back to closed unexpectedly.  The Ruby directory is readable by the dll.  Followed the instructions to a T and keep getting this.  I would love to get this working but something tells me it wont be till at least a beta comes out before I could get it to work.
    Monday, March 26, 2007 5:17 PM