The Internal Server message means that the scheduler could not start. You should check if you can reach the cgi URL with a browser and check the apache log file why the error 500 is happening. Seems like a configuration problem with apache.
Regards Christian Am 03.05.2014 20:07, schrieb Nicolas Rakover: > Thanks for the quick responses. > > Recompiling fixed the issue with shmem, and I no longer see any errors in > my scheduler logs. However, work units are still not being delivered to > clients. In the client message logs I get the following every time I click > on "Update" in the BOINC Manager: > > Sat May 3 12:57:08 2014 | full | update requested by user > Sat May 3 12:57:09 2014 | full | Sending scheduler request: Requested by > user. > Sat May 3 12:57:09 2014 | full | Requesting new tasks for CPU and intel_gpu > Sat May 3 12:57:11 2014 | full | Scheduler request failed: HTTP internal > server error > > (where "full" is just the project name) > > I know this is a sort of generic error message that implies that the issue > lies in the server, but when I look at my log files I don't see anything > unusual. The only strange thing is that scheduler.log doesn't say anything > new. I've added a bunch of debug entries to my config.xml but I get no new > logs out of the scheduler. > > Any suggestions where to dig next? > > -Nico > > > On Fri, May 2, 2014 at 12:13 PM, Jon Sonntag <[email protected]> wrote: > >> If the issue turns out to be that you don't have enough shared memory (e.g. >> feeder size > 200) you do do not need to recompile the kernel or even >> restart the server with recent Debian versions. I found the instructions >> which worked well for me at >> >> https://www.msi.umn.edu/~cpsosa/ChemApps/QuaChem/gaussian_tech/support/download/Linux_shared_mem_seg.html >> >> >> Jon Sonntag >> >> >> On Fri, May 2, 2014 at 12:45 AM, David Anderson <[email protected] >>> wrote: >> >>> Try recompiling everything; do >>> make clean ; make >>> in your source tree, then run the "upgrade" script: >>> http://boinc.berkeley.edu/trac/wiki/ToolUpgrade >>> Let me know if any problems. >>> >>> -- David >>> >>> >>> On 01-May-2014 8:20 PM, Nicolas Rakover wrote: >>> >>>> Hi all, >>>> >>>> I am building my own BOINC project and are having problems with getting >> my >>>> generated work units delivered to volunteers. >>>> >>>> I think I've identified the problem by inspecting >>>> log_debian6/scheduler.log , >>>> where I find the following message: >>>> >>>> Error in structure: shmem >>>> >>>> 2014-05-01 23:04:37.0288 [PID=2096 ] [CRITICAL] shmem has wrong struct >>>> sizes - recompile >>>> >>>> >>>> Additionally, trying to run the command bin/show_shmem results in the >>>> following output: >>>> >>>> shmget in attach_shmem: No such file or directory >>>> >>>> can't attach shmem: key <some key value> >>>> >>>> >>>> Any ideas what is going on? Let me know if you need any additional >>>> information. >>>> >>>> >>>> Thanks, >>>> >>>> Nico >>>> >>>> >>>> _______________________________________________ >>> boinc_dev mailing list >>> [email protected] >>> http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev >>> To unsubscribe, visit the above URL and >>> (near bottom of page) enter your email address. >>> >> _______________________________________________ >> boinc_dev mailing list >> [email protected] >> http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev >> To unsubscribe, visit the above URL and >> (near bottom of page) enter your email address. >> > > > _______________________________________________ boinc_dev mailing list [email protected] http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
