[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] Re: [Globus-discuss] problem with globus and condor

Hi Mats,

I appreciate your dedication to helping me......

I am using the nmi version 5.1 which has globus 3.2 in it and condor 6.6.6. I downloaded condor 6.6.8 and installed that hoping it would help. I have not applied any patches at all and havent seen any to apply to be honest. I am not running NFS at all on these machines. In fact I am only testing on one machine at the present time as I thought it would be better to get it running before putting the other machines in the mix. So I am starting the condor daemons on the same machine that I am submitting the job and using my own id on that same machine to submit a job to it.


Mats Rynge wrote:

* Darryl Cook <dlc@xxxxxxxxxxxxxxx> [2005-02-07 15:53:51 -0500]:

I went back and recompiled *everything* and reinstalled everything. The globus-run-job does now give me the date and time like you would expect. But when I run condor I still dont get anything and really dont see any errors either. The job goes in the queue and initially is Idle ....then it runs and no output to be seen. The log file says:

I have lost track of what version of Globus you are using, but one longstanding GRAM problem has been race condition problems with NFS filesystems.

What version of Globus are you using? Have you applied any patches?

Is the home directory of the user who is running the jobs on NFS?