From 8e11576524032345a2c2320b06a22d5f926d054f Mon Sep 17 00:00:00 2001 From: James Corbett Date: Wed, 29 Apr 2026 12:41:17 -0700 Subject: [PATCH 1/5] Fix typos and formatting issues in documentation - Fix grammar: 'an Flux' -> 'a Flux' in jobscripts README - Fix typo: 'in sufficient time' -> 'insufficient time' in commands README - Fix typo: 'scipts' -> 'scripts' in commands README - Fix formatting: add spaces after backticks in resmgrs README - Fix filename: 'resource manager.py' -> 'resourcemanager.py' in resmgrs README - Fix Docker image name: 'scr/SCR:v' -> 'libscr/scr:v' in docker README Co-Authored-By: Claude Sonnet 4 --- scripts/jobscripts/README.md | 2 +- scripts/python/commands/README.md | 4 ++-- scripts/python/scrjob/resmgrs/README.md | 14 +++++++------- src/test/docker/README.md | 2 +- 4 files changed, 11 insertions(+), 11 deletions(-) diff --git a/scripts/jobscripts/README.md b/scripts/jobscripts/README.md index f20530b2..a8c4ac4a 100644 --- a/scripts/jobscripts/README.md +++ b/scripts/jobscripts/README.md @@ -34,4 +34,4 @@ Before the script exits, it runs ``scr_postrun`` to scavenge any cached dataset. - ``scr_srun_loop.sh`` - Launch with ``srun`` in a SLURM allocation - ``scr_jsrun_loop.sh`` - Launch with ``jsrun`` in an LSF allocation -- ``scr_flux_run_loop.sh`` - Launch with ``flux run`` in an Flux allocation +- ``scr_flux_run_loop.sh`` - Launch with ``flux run`` in a Flux allocation diff --git a/scripts/python/commands/README.md b/scripts/python/commands/README.md index 25c78b37..5e830a53 100644 --- a/scripts/python/commands/README.md +++ b/scripts/python/commands/README.md @@ -9,12 +9,12 @@ Detailed usage for the scripts in this directory is provided in the SCR user doc - ``scr_prerun`` - Execute before the first SCR job in an allocation - ``scr_postrun`` - Execute after the final SCR job in an allocation; scavenges any cached datasets - ``scr_list_down_nodes`` - Reports list of currently failed nodes in an allocation, if any -- ``scr_should_exit`` - Indicates whether one should stop launching SCR runs within an allocation; checks for active halt condition, insufficient nodes, or in sufficient time +- ``scr_should_exit`` - Indicates whether one should stop launching SCR runs within an allocation; checks for active halt condition, insufficient nodes, or insufficient time - ``scr_halt`` - View/edit/remove conditions in the halt file ## scr\_run -The ``scr_run`` script provides a high-level wrapper around the above scipts. +The ``scr_run`` script provides a high-level wrapper around the above scripts. It can automatically relaunch a job and avoid down nodes after detecting a failure, and it scavenges any cached datasets before exiting the allocation. diff --git a/scripts/python/scrjob/resmgrs/README.md b/scripts/python/scrjob/resmgrs/README.md index 595bebe5..36ad9940 100644 --- a/scripts/python/scrjob/resmgrs/README.md +++ b/scripts/python/scrjob/resmgrs/README.md @@ -4,15 +4,15 @@ like acquiring the allocation id, the list of compute nodes, and the expected end time of the allocation. Base class: -- ``resource manager.py`` - Defines the ``ResourceManager`` base class that each resource manager implements -- ``auto.py``- Defines the ``AutoResourceManager`` class that instantiates a resource manager depending on the system environment +- ``resourcemanager.py`` - Defines the ``ResourceManager`` base class that each resource manager implements +- ``auto.py`` - Defines the ``AutoResourceManager`` class that instantiates a resource manager depending on the system environment Existing resource manager classes: -- ``flux.py``- Flux -- ``lsf.py``- IBM LSF -- ``pbsalps.py``- Cray Torque PBS with ALPS launcher -- ``pmix.py``- PMIX (not functional) -- ``slurm.py``- SLURM +- ``flux.py`` - Flux +- ``lsf.py`` - IBM LSF +- ``pbsalps.py`` - Cray Torque PBS with ALPS launcher +- ``pmix.py`` - PMIX (not functional) +- ``slurm.py`` - SLURM # Adding a new resource manager diff --git a/src/test/docker/README.md b/src/test/docker/README.md index 367ddf02..d41c9c9e 100644 --- a/src/test/docker/README.md +++ b/src/test/docker/README.md @@ -7,7 +7,7 @@ of SCR. Docker is used under CI to speed up deployment of an environment with correct build dependencies and to keep a docker image deployed at `libscr/scr` DockerHub with latest master build -(`libscr/scr:latest`) and tagged builds (`scr/SCR:v`), +(`libscr/scr:latest`) and tagged builds (`libscr/scr:v`), which can be used by other projects to build against the latest or a tagged version of SCR. From 61abaed2236ca7f2286b442676eb0edfa9344b33 Mon Sep 17 00:00:00 2001 From: James Corbett Date: Wed, 29 Apr 2026 12:46:50 -0700 Subject: [PATCH 2/5] Fix heading level in nodetests README Change "## Adding new node health tests" to "# Adding new node health tests" to match the heading structure used in other scrjob READMEs (launchers, resmgrs, remoteexec). Co-Authored-By: Claude Sonnet 4 --- scripts/python/scrjob/nodetests/README.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/scripts/python/scrjob/nodetests/README.md b/scripts/python/scrjob/nodetests/README.md index 3ad75321..ac759e0e 100644 --- a/scripts/python/scrjob/nodetests/README.md +++ b/scripts/python/scrjob/nodetests/README.md @@ -21,7 +21,7 @@ Existing tests: - ``echo.py`` - Nodes that fail to execute an ``echo UP`` command - ``dir_capacity.py`` - Nodes that fail the ``scr_check_node.py`` test, which verifies that cache and control directories are writable and optionally have a minimum capacity -## Adding new node health tests +# Adding new node health tests The steps to add a new node test are described below. From 0aa1ed0c4e43c2964d2d12e1b710a3a8875b1e3b Mon Sep 17 00:00:00 2001 From: James Corbett Date: Wed, 29 Apr 2026 12:47:05 -0700 Subject: [PATCH 3/5] Fix formatting in Python scripts README - Add missing space: "ClusterShell``NodeSet``" -> "ClusterShell ``NodeSet``" - Add protocol to URL: clustershell.readthedocs.io -> https://clustershell.readthedocs.io Co-Authored-By: Claude Sonnet 4 --- scripts/python/README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/scripts/python/README.md b/scripts/python/README.md index 19c172ae..2032ccb6 100644 --- a/scripts/python/README.md +++ b/scripts/python/README.md @@ -48,6 +48,6 @@ Node groups can be bound together by differing lists. Library defaults may need to be overridden for identifying nodes (see bottom of the config.html link) -The ClusterShell``NodeSet`` class supports more operations than ``scr_hostlist``. +The ClusterShell ``NodeSet`` class supports more operations than ``scr_hostlist``. See the NodeSet class: ``ClusterShell.NodeSet.NodeSet``. -Using package: clustershell.readthedocs.io/en/latest/guide/taskmgnt.html +Using package: https://clustershell.readthedocs.io/en/latest/guide/taskmgnt.html From 250a89c7d78575f0818c265a122cae7e782c17d0 Mon Sep 17 00:00:00 2001 From: James Corbett Date: Wed, 29 Apr 2026 12:51:02 -0700 Subject: [PATCH 4/5] Fix typos in RST documentation - doc/rst/users/quick.rst: 'direcotry' -> 'directory' - doc/rst/users/build.rst: 'defalts' -> 'defaults' - doc-dev/rst/developers/hash.rst: "it's value" -> "its value" Co-Authored-By: Claude Sonnet 4 --- doc-dev/rst/developers/hash.rst | 2 +- doc/rst/users/build.rst | 2 +- doc/rst/users/quick.rst | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/doc-dev/rst/developers/hash.rst b/doc-dev/rst/developers/hash.rst index 6907343d..782aa1bd 100644 --- a/doc-dev/rst/developers/hash.rst +++ b/doc-dev/rst/developers/hash.rst @@ -365,7 +365,7 @@ one could do the following: int64_t current_timestep = -1; if (scr_hash_util_get_int64(hash, "TIMESTEP", ¤t_timestep) == SCR_SUCCESS) { - /* TIMESTEP was set, and it's value is now in current_timestep */ + /* TIMESTEP was set, and its value is now in current_timestep */ } else { /* TIMESTEP was not set, and current_timestep is still -1 */ } diff --git a/doc/rst/users/build.rst b/doc/rst/users/build.rst index a981459c..4d7caa2f 100644 --- a/doc/rst/users/build.rst +++ b/doc/rst/users/build.rst @@ -89,7 +89,7 @@ One can disable portions of the SCR build if they are not needed: * :code:`-DENABLE_IBM_BBAPI=[ON/OFF]` : Whether to enable IBM Burst Buffer support for file transfers, defaults to :code:`OFF` * :code:`-DENABLE_CRAY_DW=[ON/OFF]` : Whether to enable Cray DataWarp support for file transfers, defaults to :code:`OFF` -* :code:`-DENABLE_PDSH=[ON/OFF]` : Whether to use pdsh to check node health and scavenge files, defalts to :code:`ON` +* :code:`-DENABLE_PDSH=[ON/OFF]` : Whether to use pdsh to check node health and scavenge files, defaults to :code:`ON` * :code:`-DBUILD_PDSH=[ON/OFF]`: CMake can automatically download and build the PDSH dependency, defaults to :code:`OFF` * :code:`-DWITH_PDSH_PREFIX=[path to PDSH]`: Path to an existing PDSH installation (should not be used with :code:`BUILD_PDSH`) diff --git a/doc/rst/users/quick.rst b/doc/rst/users/quick.rst index 0acf3658..bfc56747 100644 --- a/doc/rst/users/quick.rst +++ b/doc/rst/users/quick.rst @@ -79,7 +79,7 @@ within the CMake :code:`build` directory: cd examples -If you still have this direcotry, +If you still have this directory, then skip ahead to the next section to run :code:`test_api.c`. Alternatively, you will find source files for example programs From b6a23ca8bd6a00ef6d2dfd42d79678dc901ab316 Mon Sep 17 00:00:00 2001 From: James Corbett Date: Wed, 29 Apr 2026 12:52:33 -0700 Subject: [PATCH 5/5] Fix errors in developer RST documentation - overview.rst: Fix format string syntax in code example - flow_api.rst: Fix duplicate word 'the the' -> 'the' Co-Authored-By: Claude Sonnet 4 --- doc-dev/rst/developers/flow_api.rst | 2 +- doc-dev/rst/developers/overview.rst | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/doc-dev/rst/developers/flow_api.rst b/doc-dev/rst/developers/flow_api.rst index 20291316..5a0c0614 100644 --- a/doc-dev/rst/developers/flow_api.rst +++ b/doc-dev/rst/developers/flow_api.rst @@ -442,7 +442,7 @@ LOOP ROUNDS #. If we need to send files to our self, just move (rename) each file, update the filemap, and loop to the next round. -#. Otherwise, if we have files for this round but the the owner rank +#. Otherwise, if we have files for this round but the owner rank does not need them, delete them. #. If we do not need to send or receive any files this round, loop to diff --git a/doc-dev/rst/developers/overview.rst b/doc-dev/rst/developers/overview.rst index 695de317..2fbb96bd 100644 --- a/doc-dev/rst/developers/overview.rst +++ b/doc-dev/rst/developers/overview.rst @@ -56,7 +56,7 @@ following: /* build file name of checkpoint file for this rank */ char checkpoint_file[256]; - sprintf(checkpoint_file, "%s/rank_%d".ckpt", + sprintf(checkpoint_file, "%s/rank_%d.ckpt", checkpoint_dir, rank ); @@ -118,7 +118,7 @@ Each change is numbered for further discussion below. /* build file name of checkpoint file for this rank */ char checkpoint_file[256]; - sprintf(checkpoint_file, "%s/rank_%d".ckpt", + sprintf(checkpoint_file, "%s/rank_%d.ckpt", checkpoint_dir, rank );