Changelog ========= Development version ------------------- 0.8.2 / 2023-06-15 ------------------ - Extend OARCluster implementation to let OAR take into account the memory parameter (:pr:`598`, :pr:`595`) 0.8.1 / 2022-10-04 ------------------ - Fix the multiple ``--interface`` CLI argument bug (:pr:`591`) - Change ``project`` to ``account`` where appropriate (PBS, Slurm) (:pr:`586`) - Do not skip ``job_extra_directives`` with ``header_skip`` values and rename ``header_skip`` to ``job_directives_skip`` (:pr:`584`) - Various CI updates, HTCondor Docker build (:pr:`588`, :pr:`587`, :pr:`583`, :pr:`582`, :pr:`581`, :pr:`580`) Thanks to `@jolange `_ and `@guillaumeeb `_. 0.8.0 / 2022-08-29 ------------------ - Use --nworkers instead of deprecated --nprocs in the generated job scripts (:pr:`560`) - Drop support for Python 3.7 (:pr:`562`) - Rename ``env_extra`` kwarg to ``job_script_prologue`` (:pr:`575`) - Rename ``extra`` kwarg to ``worker_extra_args`` (:pr:`576`) - Rename ``job_extra`` kwarg to ``job_extra_directives`` (:pr:`577`) - Fixing CI failures (:pr:`562`, :pr:`574`) - Fix behaviour of env_extra for HTCondor and other related fixes (:pr:`563`, :pr:`570`, :pr:`572`) - Add batch_name to match the name of the Dask worker in HTCondor (:pr:`571`) Thanks to `@jolange `_, `@ikabadzhov `_ and `@guillaumeeb `_. 0.7.4 / 2022-07-13 ------------------ - Testing fixes (:pr:`538`, :pr:`537`, :pr:`533`, :pr:`532`, :pr:`531`, :pr:`523`, :pr:`511`) - Drop support for Python 3.6 (:pr:`279`) - Fix docstring as HTCondor needs no shared filesystem (:pr:`536`) - Fix some utils deprecations (:pr:`529`) - Add the possibility to use TLS and auto generate certificates (:pr:`519`, :pr:`524`, :pr:`526`, :pr:`527`) - Adding extra argument to condor_submit (:pr:`411`) 0.7.3 / 2021-07-22 ------------------ - Override _new_worker_name to make it easier to use job arrays (:pr:`480`) - Drop support for Python 3.5 (:pr:`456`) - Remove `FutureWarning`s from dask utils functions. (:pr:`503` and :pr:`506`) 0.7.2 / 2020-12-07 ------------------ - Use ``Status`` enum (:pr:`476`) - Bump GHA ``setup-miniconda`` version (:pr:`474`) - Build docker images for scheduled runs (:pr:`468`) - Blacken after likely a black change - Add GH action to periodically build docker images (:pr:`455`) - Fix link format inside an italicised text (:pr:`460`) - ``MoabCluster``: fix bug where ``MoabCluster`` was using the ``jobqueue.pbs`` config section rather than the ``jobqueue.moab`` section (:pr:`450`) - Updating ``start_workers`` to scale in examples (:pr:`453`) - Fixing typo in ``core.py`` (:pr:`454`) - Improve doc about GiB vs GB - Fix math typo in GiB definition (:pr:`445`) - Improve doc about customising dashboard link - Remove Travis mentions following Github Actions switch (:pr:`444`) - Improve error message. - Tweak name in ``cluster.job_script()`` (:pr:`439`) - Switch from Travis to GitHub Actions (:pr:`435`) - All cluster classes: fix a bug that would allow to pass any named parameter without an error (:pr:`398`) - Use pre-built docker images to speed up CI (:pr:`432`) - Rename common work-arounds section. - Kick-off doc section about common work-arounds (:pr:`430`) - Clean up parametrized tests (:pr:`429`) - All cluster classes: ``scheduler_options`` parameter can be set through the config file in the ``scheduler-options`` section (:pr:`405`) - Add minimal HTCondor CI support (:pr:`420`) - Add content about the python executable used by workers in SLURM (:pr:`409`) - Remove ``config_name`` from cluster classes (:pr:`426`) - Fix mysql version to get Slurm CI green (:pr:`423`) - Fix URL for miniconda download (:pr:`412`) 0.7.1 / 2020-03-26 ------------------ - all cluster classes: add ``scheduler_options`` allows to pass parameters to the Dask scheduler. For example ``scheduler_options={'interface': 'eth0', dashboard_addresses=':12435')`` (:pr:`384`). Breaking change: using ``port`` or ``dashboard_addresses`` arguments raises an error. They have to be passed through ``scheduler_options``. - all cluster classes: ``processes`` parameter default has changed. By default, ``processes ~= sqrt(cores)`` so that the number of processes and the number of threads per process is roughly the same. Old default was to use one process and only threads, i.e. ``proccesses=1``, ``threads_per_process=cores``. (:pr:`375`) - all cluster classes: ``interface`` was ignored when set in a config file. (:pr:`366`) - ``LSFCluster``: switch to ``use_stdin=True`` by default (:pr:`388`). - ``LSFCluster``: add ``use_stdin`` to ``LSFCluster``. This switches between ``bsub < job_script`` and ``bsub job_script`` to launch a ``LSF`` job (:pr:`360`). - ``HTCondorCluster``: support older ``HTCondor`` versions without ``-file`` argument (:pr:`351`). - ``OARCluster``: fix bug (forgotten async def) in ``OARCluster._submit_job`` (:pr:`380`). 0.7.0 / 2019-10-09 ------------------ - Base Dask-Jobqueue on top of the core ``dask.distributed.SpecCluster`` class (:pr:`307`) This is nearly complete reimplementation of the dask-jobqueue logic on top of more centralized logic. This improves standardization and adds new features, but does include the following **breaking changes**: + The ``cluster.start_workers`` method has been removed. Use ``cluster.scale`` instead. + The ``cluster.stop_all_jobs()`` method has been removed. Please use ``cluster.scale(0)`` instead. + The attributes ``running_jobs``, ``pending_jobs``, and ``cancelled_jobs`` have been removed. These have been moved upstream to the ``dask.distributed.SpecCluster`` class instead as ``workers`` and ``worker_spec``, as well as ``.plan``, ``.requested``, and ``.observed``. + The ``name`` attribute has been moved to ``job_name``. - You can now specify jobs in ``.scale`` and ``.adapt``: for example ``cluster.scale(jobs=2)`` and ``cluster.adapt(minimum_jobs=0, maximum_jobs=10)``. Specifying scaling in terms of jobs is generally more intuitive than in terms of Dask workers. This was part of :pr:`307`. - Update ``.scale()`` and ``.adapt()`` docstrings (:pr:`346`) - Update interactive docs (:pr:`340`) - Improve error message when cores or memory is not specified (:pr:`331`) - Fix Python 3.5.0 support in setup.py (:pr:`317`) 0.6.3 / 2019-08-18 ------------------ - Compatibility with Dask 2.3.0: add scheduler_info from local_cluster (:pr:`313`) - Remove lingering Python 2 specific code (:pr:`308`) - Remove __future__ imports since we depend on Python >3.5 (:pr:`311`) - Remove Python 3 check for black in CI (:pr:`315`) 0.6.2 / 2019-07-31 ------------------ - Ensure compatibility with Dask 2.2 (:pr:`303`) - Update documentation 0.6.1 / 2019-07-25 ------------------ - more fixes related to ``distributed >= 2`` changes (:pr:`278`, :pr:`291`) - ``distributed >= 2.1`` is now required (:pr:`295`) - remove deprecated ``threads`` parameter from all the ``Cluster`` classes (:pr:`297`) - doc improvements (:pr:`290`, :pr:`294`, :pr:`296`) 0.6.0 / 2019-07-06 ------------------ - Drop Python 2 support (:pr:`284`) - Fix adaptive compatibility with SpecificationCluster in Distributed 2.0 (:pr:`282`) 0.5.0 / 2019-06-20 ------------------ - Keeping up to date with Dask and Distributed (:pr:`268`) - Formatting with Black (:pr:`256`, :pr:`248`) - Improve some batch scheduler integration (:pr:`274`, :pr:`256`, :pr:`232`) - Add HTCondor compatibility (:pr:`245`) - Add the possibility to specify named configuration (:pr: `204`) - Allow free configuration of Dask diagnostic_port (:pr: `192)` - Start work on ClusterManager, see https://github.com/dask/distributed/issues/2235 (:pr:`187`, :pr:`184`, :pr:`183`) - A lot of other tiny fixes and improvements(:pr:`277`, :pr:`261`, :pr:`260`, :pr:`250`, :pr:`244`, :pr:`200`, :pr:`189`) 0.4.1 / 2018-10-18 ------------------ - Handle worker restart with clearer message (:pr:`138`) - Better error handling on job submission failure (:pr:`146`) - Fixed Python 2.7 error when starting workers (:pr:`155`) - Better handling of extra scheduler options (:pr:`160`) - Correct testing of Python 2.7 compatibility (:pr:`154`) - Add ability to override python used to start workers (:pr:`167`) - Internal improvements and edge cases handling (:pr:`97`) - Possibility to specify a folder to store every job logs file (:pr:`145`) - Require all cores on the same node for LSF (:pr:`177`) 0.4.0 / 2018-09-06 ------------------ - Use number of worker processes as an argument to ``scale`` instead of number of jobs. - Bind scheduler bokeh UI to every network interfaces by default. - Adds an OAR job queue system implementation. - Adds an LSF job queue system implementation. - Adds some convenient methods to JobQueueCluster objects: ``__repr__``, ``stop_jobs()``, ``close()``.