AdHocProvider(channels=, worker_init='', cmd_timeout=30, parallelism=1, move_files=None)¶
Ad-hoc execution provider
This provider is used to provision execution resources over one or more ad hoc nodes that are each accessible over a Channel (say, ssh) but otherwise lack a cluster scheduler.
channels (list of Channel ojects) – Each channel represents a connection to a remote node
worker_init (str) – Command to be run before starting a worker, such as ‘module load Anaconda; source activate env’. Since this provider calls the same worker_init across all nodes in the ad-hoc cluster, it is recommended that a single script is made available across nodes such as ~/setup.sh that can be invoked.
cmd_timeout (int) – Duration for which the provider will wait for a command to be invoked on a remote system. Defaults to 30s
parallelism (float) – Determines the ratio of workers to tasks as managed by the strategy component
__init__(channels=, worker_init='', cmd_timeout=30, parallelism=1, move_files=None)¶
Initialize self. See help(type(self)) for accurate signature.
__init__([channels, worker_init, …])
Cancel a list of jobs with job_ids
Get status of the list of jobs with job_ids
submit(command, tasks_per_node[, job_name])
Submits the command onto a channel from the list of channels
Number of cores to provision per node.
Provides the label for this provider
Real memory to provision per node in GB.
Returns the interval, in seconds, at which the status method should be called.