v1.0.0

SlurmClusterManager v1.0.0

Diff since v0.1.5

Breaking changes

• verbose field removed from SlurmManager. The messages that were previously generated under this option are now logged using @debug.

Merged pull requests:

• [BREAKING] Remove the verbose field, and use the Julia logging system instead (#23) (@DilumAluthge)
• Test suite: Increase the timeout to 10 minutes; also, allow customizing the timeout with JULIA_SLURMCLUSTERMANAGER_TEST_TIMEOUT_SECONDS (#39) (@DilumAluthge)

v0.1.5

SlurmClusterManager v0.1.5

Diff since v0.1.4

Merged pull requests:

• Test suite: Only run the c1,c2 hostname test in CI (#29) (@DilumAluthge)
• Print a @warn message if the user provides an unsupported kwarg (#40) (@DilumAluthge)

v0.1.4

SlurmClusterManager v0.1.4

Diff since v0.1.3

Merged pull requests:

• By default, propagate Base.active_project(), Base.LOAD_PATH, and Base.DEPOT_PATH to the workers (but allow the user to override that behavior by specifying JULIA_{PROJECT,LOAD_PATH,DEPOT_PATH} in params[:env]) (#30) (@DilumAluthge)
• Don't use bare using Foo (#36) (@DilumAluthge)
• Test suite: Change some println() statements to @info statements (#38) (@DilumAluthge)

Closed issues:

• addprocs does not activate workers in current project (#16)

v0.1.3

SlurmClusterManager v0.1.3

Diff since v0.1.2

Merged pull requests:

• Add support for params[:env] (that is, the env kwarg to the addprocs() function) (#20) (@DilumAluthge)
• CI: Update the Julia versions that are used in CI (#21) (@DilumAluthge)
• If an error is encountered, print the log message with @error (instead of println) (#24) (@DilumAluthge)
• CI: Change docker-compose to docker compose (#25) (@DilumAluthge)
• CI: Multiple tweaks to the CI workflow file (#26) (@DilumAluthge)
• Add some more logging (#27) (@DilumAluthge)
• Test suite: Change @asserts to errors, and change bare using Foo to using Foo: name, anothername, ... (#28) (@DilumAluthge)
• CI: In the Dockerfile, pin the full sha256 digest in the FROM line (#32) (@DilumAluthge)
• switch to updated ghcr image (#33) (@kleinhenz)
• Put the entire test suite inside a top-level @testset (#34) (@DilumAluthge)
• More test suite tweaks (#35) (@DilumAluthge)

Closed issues:

• Shebang pass options in scripts (#9)
• How to set the value JULIA_WORKER_TIMEOUT=300 (#12)
• Failed to addprocs (#15)
• Is there a GitHub repo for the jkleinh/slurm-cluster Docker image? (#31)

v0.1.2

SlurmClusterManager v0.1.2

Diff since v0.1.1

Closed issues:

• Parsing error (#6)
• Handling of multi-threaded computations on multiple nodes (#7)

Merged pull requests:

• Pass cluster cookie to workers as std input rather than on command line (#8) (@mattwigway)

v0.1.1

SlurmClusterManager v0.1.1

Closed issues:

• ntasks and jobid (#1)
• Bad resolution of relative path in include statement (#2)
• n allocated processes, and n+1 spawned in julia (#3)
• Output file is not populated in real time (#4)