You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In Worker::run_one , it's possible to specify ID of a worker that does not exists. It's also possible to build a "workerless" worker, i.e. only coordinator without any workers, with WorkerBuilder::workers(0). Both cases are demonstrated here.
We may also want to introduce a dedicated error variant and check if the worker ID (index) they are specifying in within the range OR just disallow picking a worker in Worker::run_one.
As for a "workerless" worker, we may want to substitute WorkerBuilder::workers for WorkerBuilder::extra_workers so that it's a at least a coordinator plus one worker plus extra workers (which will default to 0).
The text was updated successfully, but these errors were encountered:
As a quick fix would be to take a 'worker_id mod workers_count' in 'Worker::run_one' to prevent out of range indexing, but I think one of the approaches from above could go into release 0.14.
Other things that may be batched are:
dependencies bump (there is a PR that's on hold for now)
we may want 'Job::retry' to be an Option for consistency (see 'Failure::retry_count')
In
Worker::run_one
, it's possible to specify ID of a worker that does not exists. It's also possible to build a "workerless" worker, i.e. only coordinator without any workers, withWorkerBuilder::workers(0)
. Both cases are demonstrated here.We may also want to introduce a dedicated error variant and check if the worker ID (index) they are specifying in within the range OR just disallow picking a worker in
Worker::run_one
.As for a "workerless" worker, we may want to substitute
WorkerBuilder::workers
forWorkerBuilder::extra_workers
so that it's a at least a coordinator plus one worker plus extra workers (which will default to 0).The text was updated successfully, but these errors were encountered: