Newer
Older
This file describes changes in recent versions of SLURM. It primarily
documents those changes that are of interest to users and admins.
* Changes in SLURM 1.3.0-pre8
=============================
-- Modify how strings are packed in the RPCs, Maximum string size
increased from 64KB (16-bit size field) to 4GB (32-bit size field).
-- Fix bug that prevented time value of "INFINITE" from being processed.
-- Added new srun/sbatch option "--open-mode" to control how output/error
files are opened ("t" for truncate, "a" for append).
-- Added checkpoint/xlch plugin for use with XLCH (Hongjia Cao, NUDT).
-- Added srun option --checkpoint-path for use with XLCH (Hongjia Cao, NUDT).
-- Added new srun/salloc/sbatch option "--acctg-freq" for user control over
accounting data collection polling interval.
* Changes in SLURM 1.3.0-pre7
=============================
-- Fix a bug in the processing of srun's --exclusive option for a job step.
* Changes in SLURM 1.3.0-pre6
=============================
-- Add support for configurable number of jobs to share resources using the
partition Shared parameter in slurm.conf (e.g. "Shared=FORCE:3" for two
jobs to share the resources). From Chris Holmes, HP.
-- Made salloc use api instead of local code for message handling.
* Changes in SLURM 1.3.0-pre5
=============================
-- Add select_g_reconfigure() function to node changes in slurmctld configuration
that can impact node scheduling.
-- scontrol to set/get partition's MaxTime and job's Timelimit in minutes plus
new formats: min:sec, hr:min:sec, days-hr:min:sec, days-hr, etc.
-- scontrol "notify" command added to send message to stdout of srun for
specified job id.
Moe Jette
committed
-- For BlueGene, make alpha part of node location specification be case insensitive.
-- Report scheduler-plugin specific configuration information with the
"scontrol show configuration" command on the SCHEDULER_CONF line. This
information is not found in the "slurm.conf" file, but a scheduler plugin
specific configuration (e.g. "wiki.conf").
-- sview partition information reported now includes partition priority.
-- Expand job dependency specification to support concurrent execution,
testing of job exit status and multiple job IDs.
* Changes in SLURM 1.3.0-pre4
=============================
-- Job step launch in srun is now done from the slurm api's all further
modifications to job launch should be done there.
-- Add new partition configuration parameter Priority. Add job count to
Shared parameter.
-- Add new configuration parameters DefMemPerTask, MaxMemPerTask, and
SchedulerTimeSlice.
-- In sched/wiki2, return REJMESSAGE with details on why a job was
requeued (e.g. what node failed).
* Changes in SLURM 1.3.0-pre3
=============================
-- Added srun option "--checkpoint=time" for job step to automatically be
checkpointed on a period basis.
-- Change behavior of "scancel -s KILL <jobid>" to send SIGKILL to all job
steps rather than cancelling the job. This now matches the behavior of
all other signals. "scancel <jobid>" still cancels the job and all steps.
-- Add support for new job step options --exclusive and --immediate. Permit
job steps to be queued when resources are not available within an existing
job allocation to dedicate the resources to the job step. Useful for
executing simultaneous job steps. Provides resource management both at
the level of jobs and job steps.
srun --nodes=16 --constraint=graphics*4 ...
Based upon work by Kumar Krishna (HP, India).
-- Add multi-core options to salloc and sbatch commands (sbatch.patch and
cleanup.patch from Chris Holmes, HP).
-- In select/cons_res properly release resources allocated to job being
suspended (rmbreak.patch, from Chris Holmes, HP).
-- Removed database and jobacct plugin replaced with jobacct_storage
and jobacct_gather for easier hooks for further expansion of the
jobacct plugin.
* Changes in SLURM 1.3.0-pre2
=============================
-- Added new srun option --pty to start job with pseudo terminal attached
to task 0 (all other tasks have I/O discarded)
-- Disable user specifying jobid when sched/wiki2 configured (needed for
Moab releases until early 2007).
-- Report command, args and working directory for batch jobs with
"scontrol show job".
* Changes in SLURM 1.3.0-pre1
=============================
Christopher J. Morrone
committed
-- !!! SRUN CHANGES !!!
The srun options -A/--allocate, -b/--batch, and -a/--attach have been
removed! That functionality is now available in the separate commands
salloc, sbatch, and sattach, respectively.
-- Add new node state FAILING plus trigger for when node enters that state.
-- Add new configuration paramter "PrivateData". This can be used to
prevent a user from seeing jobs or job steps belonging to other users.
-- Added configuration parameters for node power save mode: ResumeProgram
ResumeRate, SuspendExcNodes, SuspendExcParts, SuspendProgram and
SuspendRate.
-- Slurmctld maintains the IP address (rather than hostname) for srun
communications. This fixes some possible network routing issues.
-- Added global database plugin. Job accounting and Job completion are the
first to use it. Follow documentation to add more to the plugin.
-- Removed no-longer-needed jobacct/common/common_slurmctld.c since that is
replaced by the database plugin.
-- Added new configuration parameter: CryptoType.
Moved existing digital signature logic into new plugin: crypto/openssl.
Added new support for crypto/munge (available with GPL license).
* Changes in SLURM 1.2.21
=========================
-- Fixed torque wrappers to look in the correct spot for the perl api
-- Do not treat user resetting his time limit to the current value as
an error.
-- Set correct executable names for Totalview when --multi-prog option
is used and more than one node is allocated to the job step.
-- When a batch job gets requeued, record in accounting logs that
the job was cancelled, the requeued job's submit time will be
set to the time of its requeue so it looks like a different job.
-- Prevent communication problems if the slurmd/slurmstepd have a
different JobAcct plugin configured than slurmctld.
-- Adding Gold plugin for job accounting
-- In sched/wiki2, add support for MODIFYJOB option "JOBNAME=<name>"
to modify a job's name.
-- Add configuration check for sys/syslog.h and include it as needed.
* Changes in SLURM 1.2.20
=========================
-- In switch/federation, fix small memory leak effecting slurmd.
-- Add PMI_FANOUT_OFF_HOST environment variable to control how message
forwarding is done for PMI (MPICH2). See "man srun" for details.
-- From sbatch set SLURM_NTASKS_PER_NODE when --ntasks-per-node option is
specified.
-- BLUEGENE: Documented the prefix should always be lower case and the 3
digit suffix should be uppercase if any letters are used as digits.
-- In sched/wiki and sched/wiki2, add support for --cpus-per-task option.
From Miguel Ros, BSC.
-- In sched/wiki2, prevent invalid memory pointer (and likely seg fault)
for job associated with a partition that has since been deleted.
-- In sched/wiki2 plus select/cons_res, prevent invalid memory pointer
(and likely seg fault) when a job is requeued.
-- In sched/wiki, add support for job suspend, resume, and modify.
-- In sched/wiki, add suppport for processor allocation (not just node allocation)
with layout control.
-- Prevent re-sending job termination RPC to a node that has already completed
the job. Only send it to specific nodes which have not reported completion.
-- Support larger environment variables 64K instead of BUFSIZ (8k on some
systems).
-- If a job is being requeued, job step create requests will print a
warning and repeatedly retry rather than aborting.
-- Add optional mode value to srun and sbatch --get-user-env option.
-- Print error message and retry job submit commands when MaxJobCount
is reached. From Don Albert, Bull.
-- Treat invalid begin time specification as a fatal error in sbatch and
srun. From Don Albert, Bull.
-- Validate begin time specification to avoid hours >24, minutes >59, etc.
* Changes in SLURM 1.2.19
=========================
*** NOTE IMPORTANT CHANGE IN RPM BUILD BELOW ****
-- slurm.spec file (used to build RPMs) was updated in order to support Mock, a
chroot build environment. See https://hosted.fedoraproject.org/projects/mock/
for more information. The following RPMs are no longer build by default:
aix-federation, auth_none, authd, bluegene, sgijob, and switch-elan. Change
the RPMs built using the following options in ~/rpmmacros: "%_with_authd 1",
"%_without_munge 1", etc. See the slurm.spec file for more details.
-- Print warning if non-privileged user requests negative "--nice" value on
job submission (srun, salloc, and sbatch commands).
-- In sched/wiki and sched/wiki2, add support for srun's --ntasks-per-node
option.
-- In select/bluegene with Groups defined for Images, fix possible memory
corruption. Other configurations are not affected.
-- BLUEGENE - Fix bug that prevented user specification of linux-image,
mloader-image, and ramdisk-image on job submission.
-- BLUEGENE - filter Groups specified for image not just by submitting
user's current group, but all groups the user has access to.
-- BLUEGENE - Add salloc options to specify images to be loaded (--blrts-image,
--linux-image, --mloader-image, and --ramdisk-image).
-- BLUEGENE - In bluegene.conf, permit Groups to be comma separated in addition
to colon separators previously supported.
-- sbatch will accept batch script containing "#SLURM" options and advise
changed to "#SBATCH".
-- If srun --output or --error specification contains a task number rather
than a file name, send stdout/err from specified task to srun's stdout/err
rather than to a file by the same name as the task's number.
-- For srun --multi-prog option, verify configuration file before attempting
to launch tasks, report clear explanation of any configuration file errors.
-- For sched/wiki2, add optional timeout option to srun's --get-user-env
parameter, change default timeout for "su - <user> env" from 3 to 8 seconds.
On timeout, attempt to load env from file at StateSaveLocation/env_cache/<user>.
The format of this file is the same as output of "env" command. If there
is no env cache file, then abort the request.
-- squeue modified for completing job to remove nodes that have already
completed the job before applying node filter logic.
-- squeue formatted output option added for job comment, "%q" (the obvious
choices for letters are already in use).
-- Added configure option --enable-load-env-no-login for use with Moab. If
set then the user job runs with the environment built without a login
("su <user> env" rather than "su - <user> env").
-- Fix output of "srun -o %C" (allocated CPU count) for running jobs. This was
Loading
Loading full blame...