Previous: 7. The standard universe | Index |
condor_status
, in the default version,
will not show if the nodes are compatible with the "parallel" universe,
i.e. if they are dedicated nodes connected with the dedicated scheduler.
To list the dedicated nodes, you should use some additional options:
condor_status -const '!isUndefined(DedicatedScheduler)' -format "%s\t" Name -format "%s\n" DedicatedScheduler
The output will be something containing lines as:
slot1@node-097.giunti DedicatedScheduler@node-080.giunti
DedicatedScheduler
must be the name of the node where schedd runs.
Note that it must appear exactly as printed by condor_status -schedd -format "%s\n" Machine
,
otherwise the DedicatedScheduler will not be recognized.
In the case of the gr4cloud
pool, consider that a worker node is marked as "dedicated" only
at the end of its initialization procedure, that may take some time.
If some nodes do not appear as dedicated, probably it is just because they have not completed the initialization.
If you want to run a docker universe job, you may be interested in listing the machines that are able to manage such a job. To do this, or to list the nodes that cannot run Docker jobs, use one of these commands:
condor_status -const 'hasDocker' condor_status -const '!hasDocker'
Previous: 7. The standard universe | Index |