I have been using a condor pool with 4 Windows machines that has been working without problems.
I am now trying to configure a new condor pool composed by a central manager that runs Windows and an execute machine that runs Linux.
I am running the Condor daemons in the Linux machine as root.
The software that invokes Condor requires that the submitted jobs have access to a shared directory that exists on the central manager/submitting machine.
I have therefore set up a dedicated account on the execute machine in order to run the jobs.
The execute machine has 4 cores. I have therefore defined SLOT1-4 on the following way,
SLOT1_USER = tachi-02\root
SLOT2_USER = tachi-02\root
SLOT3_USER = tachi-02\root
SLOT4_USER = tachi-02\root
DEDICATED_EXECUTE_ACCOUNT_REGEXP = True
I have then tried to store the password for this account on the execute machine
root@tachi-02:~# condor_store_cred add
Account: root@(null)
Enter password:
Operation failed.
Make sure your ALLOW_WRITE setting includes this host.
This operation worked fine on my previous Windows condor pool as long as I had the condor_schedd running when storing the password.
1. Why doesn’t it work on the Linux machine? The condor_schedd daemon is running and the variable ALLOW_WRITE is set to *.
2. Does condor_store_cred add works on a different way on Linux as compared to Windows?
3. How should I store the password on a Linux machine?
I have another question.
I have tried to set the variable DAEMON_LIST in the condor_config file of the execute machine as shown below.
## Daemons you want the master to keep running for you:
DAEMON_LIST = MASTER, STARTD
However, the variable does not seem to change even after rebooting the system.
root@tachi-02:~# condor_config_val DAEMON_LIST
COLLECTOR, MASTER, NEGOTIATOR, SCHEDD, STARTD
Why doesn’t this variable change?