Define error handling policy for jobs

Create an error handling policy that defines how to handle job failures, including retry logic, failure thresholds, and recovery strategies.

Usage

on_error(
  action = c("stop", "continue", "retry"),
  max_retries = 3,
  backoff = c("none", "linear", "exponential"),
  backoff_base = 60,
  collect_errors = TRUE
)

Arguments

action

Action to take on error: "stop", "continue", "retry"

max_retries

Maximum number of retry attempts

backoff

Backoff strategy: "none", "linear", "exponential"

backoff_base

Base delay in seconds for backoff

collect_errors

Whether to collect error messages

Value

Error policy object

Examples

# \donttest{
# Retry failed jobs up to 3 times with exponential backoff
policy <- on_error(
  action = "retry",
  max_retries = 3,
  backoff = "exponential",
  backoff_base = 60
)

# Continue on errors and collect them
policy <- on_error(
  action = "continue",
  collect_errors = TRUE
)

# Use with job submission
risky_function <- function(x) if (runif(1) > 0.5) stop("Random error") else x^2
if (Sys.which("squeue") != "") {
  jobs <- slurm_map(1:10, risky_function, .error_policy = policy)
}
# }