Re: [Condor-users] Problems with condor_g

[Alexander Dietz <dietz@xxxxxxxxxxxx>]

Hi,

I had some trouble with my certificates, but I also created a link to where the certficates are (as you describe below). Then I stopped condor (condor_off) and started it again condor_on). Then I tried again to submit jobs with condor_g, but I got the same pronlem (Failed to initialize GHAP).
Maybe it is something wrong with my certificate? How can I check this? Any other idea?

Regards
Alexander



Rod Walker wrote:

Hi,
I saw a similar problem which was due to the certificates directory not 
being found(CA's plus crls). I linked the default location, 
/etc/grid-security/certificates, to the directory specified by 
X509_CERT_DIR. But I think if you set the environement before starting 
the Scheduler then the gridmanager will respect this.

Cheers,
Rod.

On Wed, 19 Oct 2005, Alexander Dietz wrote:

  

Hi,

I have some problems to submit jobs to different pools via condor_g. I 
tried to submit either to jobmanager-condor, jobbmanager-pbs (whatever 
jobmanager is running on that pool) and also to jobmanager-fork. But all 
the jobs get hold and the only information I can get is from the log 
file which says "Failed to initialize GAHP". I tried to google that 
phrase, but got no single hit!
So what does this error mean? Could this be the reason for all of my 
jobs get hold? See below for an example of my submit file.

Regards
Alexander

Example submit-file:


    universe = globus
    executable = /bin/hostname
    globusscheduler=ligo-grid.aset.psu.edu/jobmanager-fork
    remote_initialdir=/tmp
    output= tryPSU.out
    error=tryPSU.err
    log=tryPSU.log
    copy_to_spool=false
    transfer_executable=false
    notification=NEVER
    queue