wrench::HTCondorComputeServiceProperty

class HTCondorComputeServiceProperty : public wrench::ComputeServiceProperty

Properties for an HTCondor service.

Public Static Attributes

static const std::string FCFS: Whether the HTCondorComputeService should enforce First-Come-First-Serve when scheduling jobs, or allow jobs that can run to be “backfilled” before jobs that were submitted earlier but that cannot run. Default: false.

static const std::string GRID_POST_EXECUTION_DELAY: Overhead between condor and a batch compute service for the completion of execution of grid-universe jobs. Default unit: second. Examples: “1s”, “200ms”, “1.5s”, etc.

static const std::string GRID_PRE_EXECUTION_DELAY: Overhead between condor and a batch compute service for the start of execution of grid-universe jobs. Default unit: second. Examples: “1s”, “200ms”, “1.5s”, etc.

static const std::string INSTANT_RESOURCE_AVAILABILITIES: Whether the HTCondorComputeService should contact its subordinate BareMetalComputeServices (i.e., when running non-grid jobs) to find out about their currently available resources (value “false”), or instead or instead uses the “out of simulation time” method to query these services about their currently available resources (value “true”); “true” is the default value since it reduces simulation time. The caveat that it no longer simulates the network load and overhead of all “tell me how many free resources you have right now?” and “this is what I got!” control messages.

static const std::string NEGOTIATOR_OVERHEAD: Overhead of the HTCondor Negotiator, which is invoked each time a new job is submitted or a running job completes and there are still pending jobs. Default: “0”. Default unit: second. Examples: “1s”, “200ms”, “1.5s”, etc.

static const std::string NON_GRID_POST_EXECUTION_DELAY: Overhead between condor and a bare-metal compute for the completion of execution of non-grid-universe jobs. Default unit: second. Examples: “1s”, “200ms”, “1.5s”, etc.

static const std::string NON_GRID_PRE_EXECUTION_DELAY: Overhead between condor and a bare-metal compute service for the start of execution of non-grid-universe jobs. Default unit: second. Examples: “1s”, “200ms”, “1.5s”, etc.