Daniel Smith
Daniel Smith
Automatic shutdown of a manager should occur if ~30 failed tasks in a row are found. InputError and ConvergenceError should not be counted towards this count. If a successful task...
**Describe the solution you'd like** Allow `get_collection` to query via a DOI. For example `get_collection(doi="1234:1234")`. **Describe alternatives you've considered** Querying off names works well, but alternatives searches for known data...
**Describe the bug** If a Parsl worker is killed for any reason we report this back as a task failure when it should likely be restarted. This is especially harsh...
**Describe the bug** After a failed push the managers enter an idle state that consumes core-time and does not continue with tasks. As a note this is quite rare. The...
**Is your feature request related to a problem? Please describe.** The number of running and waitings tasks on a manager is not well stated for task execution systems outside of...
**Is your feature request related to a problem? Please describe.** We have custom chunk for-based iterators throughout the code for handling multiple requests. **Describe the solution you'd like** We should...
We may be missing the transfer from a crashed service to transferring this error to the record. Need to double check this and add a test. @ChayaSt
ProcedureDataset-based collections have additional features such as `get_entry` and `get_record` that Dataset-based collections do not have. At first pass we can simply add them to Dataset derivatives, but it might...
**Describe the bug** When we query a projected procedure like so: ``` ret = client.query_procedures(id=1711898, procedure="optimization", projection={"trajectory": True, "id": True}) ``` we get the projection as expected. However if we...
**Describe the bug** The intention of the CLI with a config file was to allow overrides via the CLI if a configuration file is present. Currently we simply check if...