[ome-users] Image import problem

[Brian Ruttenberg]

Great, thanks!  At this point, I have no idea if the problem is omeis or
apache, so this should certainly help with the debug process.  I'll give
it a run and see.

As to your other questions, I'm running apache 2.2.3 on Debian.  I
haven't played around with the apache config, so almost everything
should be at the defaults.  And omeis is running locally, but the image
import is remote.  I don't suspect that would be a problem, as there are
other machines here that have been running OME for a while with no
import problems.  My guess right now would be something wrong with
apache, since I've just recently built this linux box, and the apache
version is fairly new (~Dec 06).

Again, thanks for your help!

Brian

-----Original Message-----
From: Ilya Goldberg [mailto:igg at nih.gov] 
Sent: Thursday, February 08, 2007 8:02 PM
To: Brian Ruttenberg
Cc: ome-users at lists.openmicroscopy.org.uk
Subject: Re: [ome-users] Image import problem


You might also try running the omeis torture test, found in src/C/ 
omeis/t/test-concurrent-write

This test assumes that you have the following directory structure in  
your current working directory when you execute the test (it has no  
parameters):
OMEIS-TEST
OMEIS-TEST/Files
OMEIS-TEST/Pixels

If your regular omeis repository is on a shared drive, then cd to  
your shared drive and make the above empty directory tree in your cwd.
For full effect, run it on a multi-core or multi-CPU box.  It will  
fork 64 processes and make each of them write 1000 randomly generated  
images to omeis at the same time.  It may take some time depending on  
your OS.  The images are quite small, and the omeis uniqueness  
constraint is in effect, so don't be surprised if there are many  
missing images in the final tree - this is normal.  The test will  
report any write failures to stderr.  It doesn't involve apache - its  
purely a test of the omeis code and the filesystem its using. This test
passes on local filesystems (FC4 and OS X 10.4) or NFS  
shares using Fedora 4 NFS servers and FC4 or OS X NFS clients.  YMMV.

-Ilya

On Feb 8, 2007, at 2:52 PM, Brian Ruttenberg wrote:

> Hi, I'm a relatively new OME user and I'm having some problems with
> some image importing, so I was wondering if anyone out there could  
> give me a hand...
>
> I start an OME import on a bunch of images, then after a non-
> deterministic time (the failure doesn't always happen in the same  
> place), I get this error:
>
> Importing files
>          1/296: [IN PROGRESS] Starting import
>          15/296: [IN PROGRESS] Uploaded /tmp/tmp.aMOzSN6790/125P5.TIF
>          29/296: [IN PROGRESS] Uploaded /tmp/tmp.aMOzSN6790/161NP7.TIF
>          42/296: [IN PROGRESS] Uploaded /tmp/tmp.aMOzSN6790/161P8.TIF
>          46/296: [IN PROGRESS] Uploaded /tmp/tmp.aMOzSN6790/1DMIP4.TIF
>          49/296: [IN PROGRESS] Uploaded /tmp/tmp.aMOzSN6790/1DMP12.TIF
>          54/296: [IN PROGRESS] Uploaded /tmp/tmp.aMOzSN6790/1DP42.TIF 
> Died at /usr/local/share/perl/5.8.8/OME/Image/Server.pm line 414. 
> Exiting...
>          55/296: [DIED] Uploaded /tmp/tmp.aMOzSN6790/1GVP30.TIF 
> Exiting...
>
> Checking that portion of the code, it is dying because of this:
>
>         if ($curl->status() == 200) {
>             return ($response);
>         } else {
>             die $response;
>
> So i checked the logs from the apache2 server I'm running, and it
> looks...odd.  The access log has this:
>
> ...
> 128.111.57.62 - - [08/Feb/2007:11:22:51 -0800] "POST /cgi-bin/omeis
> HTTP/1.1" 200 4 "-" "-"
> 128.111.57.62 - - [08/Feb/2007:11:22:51 -0800] "POST /cgi-bin/omeis  
> HTTP/1.1" 200 4 "-" "-"
> 128.111.57.62 - - [08/Feb/2007:11:23:07 -0800] "POST /cgi-bin/omeis  
> HTTP/1.1" 200 - "-" "-"
>
> If I'm reading this correct, apache is responding with an OK for a
> while, then on the last omeis request it responds with an OK but  
> returns no data ( the '-' after the 200 I presume).  The error.log  
> also has this:
>
> [Thu Feb 08 11:23:07 2007] [error] [client 128.111.57.62] (104)
> Connection reset by peer: ap_content_length_filter: apr_bucket_read 
> () failed
>
> which, with my limited apache experience, I'm not sure if that has
> any affect on the import.
>
> Has anyone encountered this error of similar errors before?  I've
> reinstalled OME several times, dropped the database, verified the  
> omeis executable, and I'm running out of options.  Any help would  
> be greatly appreciated.  Thank!
>
> Brian
>
> _______________________________________________
> ome-users mailing list
> ome-users at lists.openmicroscopy.org.uk
> http://lists.openmicroscopy.org.uk/mailman/listinfo/ome-users