v0.4.6

ClusterManagers v0.4.6

Diff since v0.4.5

Merged pull requests:

• enable dependabot for GitHub actions (#188) (@ranocha)
• add support for specifying slurm output file (#194) (@cnrrobertson)
• Bump actions/checkout from 2.2.0 to 4.1.1 (#195) (@dependabot[bot])
• Make SLUM worker startup more robust and provide more feedback (#200) (@oschulz)

v0.4.5

ClusterManagers v0.4.5

Diff since v0.4.4

Closed issues:

• Error launching workers: no such file or directory (#165)
• ClusterManagers can be run on top of dask clusters! (#186)

Merged pull requests:

• Update ElasticManager constructor auto IP config (#190) (@mikeingold)

v0.4.4

ClusterManagers v0.4.4

Diff since v0.4.2

Closed issues:

• [SlurmManager] 100 % CPU usage while waiting for the job to get created (#173)
• Better handling of SLURM job submission timing (#175)
• SLURM 10 nodes good, 16 nodes error (#178)
• LSF manager broken in Julia 1.8.1 (#182)

Merged pull requests:

• Add Slurm Docker tests (#154) (@simonbyrne)
• [Slurm] Add sleep while waiting for job to start (#174) (@David96)
• [Slurm] Add retry_delays when waiting for workers (#176) (@David96)
• Bump version with slurm retry_delays fix (#180) (@MilesCranmer)
• Update TagBot.yml with new template (#181) (@DrChainsaw)
• Close and reopen stream after bpeek exits (#183) (@DrChainsaw)
• improvement and bug fix for condor (#184) (@Moelf)

v0.4.2

ClusterManagers v0.4.2

Diff since v0.4.1

Closed issues:

• lsf_bpeek makes strong assumptions on iterator state of retry_delays (#171)

Merged pull requests:

• iterate over retry_delays (#172) (@DrChainsaw)

v0.4.1

ClusterManagers v0.4.1

Diff since v0.4.0

Closed issues:

• addprocs_sge works incorrectly if the work directory is not the home directory and the list of default options for qsub includes "-cwd" (#103)
• Carriage return instead of newline (#112)
• Possibly naive question: ClusterManagers work for kubernetes? (#146)
• Extra options on SGE (#151)
• Error in rmprocs SGE (#152)
• addprocs(SGEManager) fails (#158)

Merged pull requests:

• simplified qsub call, fixed kill, added working dir option + some tests (#153) (@nantonel)
• Update link to julia docs (#155) (@JonasIsensee)
• Fix up and add flexibility to HTCManager (#157) (@aminnj)
• clarify qsub flags are not limited to queue and need -q (#160) (@Moelf)
• Update README.md (#161) (@JonathanAnderson)
• Missing kwargs in qsub addproc (#162) (@nantonel)
• Add reference to K8sClusterManager in README (#166) (@omus)
• LSFManager enhancements (#167) (@tanmaykm)
• LSFManager: optionally stream worker stdout to master (#168) (@tanmaykm)
• Update patch version for tagging (#169) (@tanmaykm)

v0.4.0

ClusterManagers v0.4.0

Diff since v0.3.2

Closed issues:

• LSF manager with bsub (#72)
• Sometimes addprocs_sge works, sometimes it doesn't (#78)
• LocalAffinityManager not updated for 1.0 (#98)
• SGEManager procs failing to launch (#100)
• Error with using on remote htcondor worker (#108)
• Getting errors with Slurm (#109)
• warn function call in slurm.jl needs to be changed to a macro to be compaticle with v1 (#115)
• Slurm: Change where job files are located (#116)
• Job dies when a node fails in SLURM (#120)
• Slurm Job creation crashes on master because "SLURM_JOB_ID" not set (#127)
• Make ElasticManager print out connection instructions (#128)
• Help on PBSManager (#132)

Merged pull requests:

• add support for LSF (#74) (@bjarthur)
• Added res_list to addprocs_qrsh (#82) (@cako)
• race condition for multiple in-bound connections (#97) (@s2maki)
• fix deprecated replace call (#99) (@lstagner)
• Changes to make addprocs_sge work under julia v1.0.0 (#101) (@oschub)
• provide sockname for ElasticManager so ephemeral port can be read (#110) (@marius311)
• Update AffinityManager for Julia 1.0 (#111) (@jbrnd)
• Addressing two issues with Slurm (#119) (@mkschleg)
• Change the naming pattern of .out files to avoid conflicts (#123) (@pmarg)
• switch to Project.toml (#124) (@vchuravy)
• Newline return (#126) (@miguelraz)
• Install TagBot as a GitHub Action (#131) (@JuliaTagBot)
• refactor LSFManager to use jobs arrays (#135) (@bjarthur)
• Make ElasticManager more convenient (#136) (@marius311)
• Slurm robust to no jobid and to node warnings (#139) (@grahamas)

ClusterManagers.jl v0.3.2

Merge pull request #97 from Conning/elastic-accept-race-condition

race condition for multiple in-bound connections

ClusterManagers.jl v0.3.1

Merge pull request #95 from kaskarn/master

minor 0.7+ fixes and updates to slurm.jl

ClusterManagers.jl v0.3.0

Julia 0.7+ compatible.
Dropped support for older Julia versions.

ClusterManagers.jl v0.2.0

Merge pull request #81 from JuliaParallel/fbot/deps

Fix deprecations