flux-jobs(1)¶

SYNOPSIS¶

flux jobs [OPTIONS] [JOBID ...]

DESCRIPTION¶

flux-jobs(1) is used to list jobs run under Flux. By default only pending and running jobs for the current user are listed. Additional jobs and information can be listed using options listed below. Alternately, specific job ids can be listed on the command line to only list those job IDs.

OPTIONS¶

-a

List jobs in all states, including inactive jobs. This is shorthand for --filter=pending,running,inactive.

-A

List jobs of all users. This is shorthand for --user=all.

-n, --no-header

For default output, do not output column headers.

-u, --user=[USERNAME|UID]

List jobs for a specific username or userid. Specify all for all users.

--name=[JOB NAME]

List jobs with a specific job name.

--queue=[QUEUE]

List jobs in a specific queue.

-c, --count=N

Limit output to N jobs (default 1000)

--sinceWHEN

Limit output to jobs that have been active since a given timestamp. In other words, jobs that are currently pending, currently running, or became inactive since the given timestamp. This option implies -a if no other --filter options are specified. If WHEN begins with - character, then the remainder is considered to be a an offset in Flux standard duration (RFC 23). Otherwise, any datetime expression accepted by the Python parsedatetime module is accepted. Examples: "-6h", "-1d", "yesterday", "2021-06-21 6am", "last Monday", etc. It is assumed to be an error if a timestamp in the future is supplied.

-f, --filter=STATE|RESULT

List jobs with specific job state or result. Multiple states or results can be listed separated by comma. See JOB STATUS below for additional information. Defaults to pending,running.

-o, --format=NAME|FORMAT

Specify a named output format NAME or a format string using Python's format syntax. See OUTPUT FORMAT below for field names. Named formats may be listed via --format=help. An alternate default format can be set via the FLUX_JOBS_FORMAT_DEFAULT environment variable. Additional named formats may be registered with flux jobs via configuration. See the CONFIGURATION section for more details. A configuration snippet for an existing named format may be generated with --format=get-config=NAME.

--json

Emit data for selected jobs in JSON format. The data for multiple matching jobs is contained in a jobs array in the emitted JSON object, unless a single job was selected by jobid on the command line, in which case a JSON object representing that job is emitted on success. With -R, --recursive, each job which is also an instance of Flux will will have any recursively listed jobs in a jobs array, and so on for each sub-child.

Only the attributes which are available at the time of the flux-jobs query will be present in the returned JSON object for a job. For instance a pending job will not have runtime, waitstatus or result keys, among others. A missing key should be considered unavailable.

The --json option is incompatible with --stats and --stats-only, and any --format is ignored.

--color[=WHEN]

Control output coloring. The optional argument WHEN can be auto, never, or always. If WHEN is omitted, it defaults to always. Otherwise the default is auto.

--stats

Output a summary of job statistics before the header. By default shows global statistics. If --queue is specified, shows statistics for the specified queue. May be useful in conjunction with utilities like watch(1), e.g.:

$ watch -n 2 flux jobs --stats -f running -c 25

will display a summary of statistics along with the top 25 running jobs, updated every 2 seconds.

Note that all job failures, including canceled and timeout jobs, are collectively counted as "failed" in --stats.

--stats-only

Output a summary of job statistics and exit. By default shows global statistics. If --queue is specified, shows statistics for the specified queue. flux jobs will exit with non-zero exit status with --stats-only if there are no active jobs. This allows the following loop to work:

$ while flux jobs --stats-only; do sleep 2; done

All options other than --queue are ignored when --stats-only is used.

Note that all job failures, including canceled and timeout jobs, are collectively counted as "failed" in --stats-only.

-R, --recursive

List jobs recursively. Each child job which is also an instance of Flux is prefixed by its jobid "path" followed by the list of jobs, recursively up to any defined -L, --level. If the --stats option is used, then each child instance in the hierarchy is listed with its stats.

--recurse-all

By default, jobs not owned by the user running flux jobs are skipped with -R, --recursive, because normally Flux instances only permit the instance owner to connect. This option forces the command to attempt to recurse into the jobs of other users. Implies --recursive.

-L, --level=N

With -R, --recursive, stop recursive job listing at level N. Levels are counted starting at 0, so flux jobs -R --level=0 is equivalent to flux jobs without -R, and --level=1 would limit recursive job listing to child jobs of the current instance.

--threads=N

When flux jobs recursively queries job lists (with --recursive) or fetches info for jobs that are also instances (see instance.* fields), a pool of threads is used to parallelize the required RPCs. Normally, the default number of ThreadPoolExecutor threads is used, but by using the --threads, a specific number of threads can be chosen.

JOB STATUS¶

Jobs may be observed to pass through five job states in Flux: DEPEND, SCHED, RUN, CLEANUP, and INACTIVE (see Flux RFC 21). Under the state_single field name, these are abbreviated as D, S, R, C, and I respectively. For convenience and clarity, the following virtual job states also exist: "pending", an alias for DEPEND,SCHED; "running", an alias for RUN,CLEANUP; "active", an alias for "pending,running".

After a job has finished and is in the INACTIVE state, it can be marked with one of the possible results: COMPLETED, FAILED, CANCELED, TIMEOUT. Under the result_abbrev field name, these are abbreviated as CD, F, CA, and TO respectively.

The job status is a user friendly mix of both, a job is always in one of the following statuses: DEPEND, SCHED, RUN, CLEANUP, COMPLETED, FAILED, CANCELED, or TIMEOUT. Under the status_abbrev field name, these are abbreviated as D, S, R, C, CD, F, CA, and TO respectively.

OUTPUT FORMAT¶

The --format option can be used to specify an output format to flux-jobs(1) using Python's string format syntax. For example, the following is the format used for the default format:

{id.f58:>12} ?:{queue:<8.8} {username:<8.8} {name:<10.10+} \
{status_abbrev:>2.2} {ntasks:>6} {nnodes:>6h} \
{contextual_time!F:>8h} {contextual_info}

If a format field is preceded by the special string ?: this will cause the field to be removed entirely from output if the result would be an empty string or zero value for all jobs in the listing. E.g.:

{id.f58:>12} ?:{exception.type}

would eliminate the EXCEPTION-TYPE column if no jobs in the list received an exception. (Thus the job queue is only displayed if at least one job has a queue assigned in the default format shown above).

As a reminder to the reader, some shells will interpret braces ({ and }) in the format string. They may need to be quoted.

The special presentation type h can be used to convert an empty string, "0s", "0.0", "0:00:00", or epoch time to a hyphen. For example, normally "{nodelist}" would output an empty string if the job has not yet run. By specifying, "{nodelist:h}", a hyphen would be presented instead.

The special suffix + can be used to indicate if a string was truncated by including a + character when truncation occurs. If both h and + are being used, then the + must appear after the h.

Additionally, the custom job formatter supports a set of special conversion flags. Conversion flags follow the format field and are used to transform the value before formatting takes place. Currently, the following conversion flags are supported by flux-jobs:

!D: convert a timestamp field to ISO8601 date and time (e.g. 2020-01-07T13:31:00). Defaults to empty string if timestamp field does not exist.
!d: convert a timestamp to a Python datetime object. This allows datetime specific format to be used, e.g. {t_inactive!d:%H:%M:%S}. Additionally, width and alignment can be specified after the time format by using two colons (::), e.g. {t_inactive!d:%H:%M:%S::>20}. Returns an empty string (or "-" if the h suffix is used) for an unset timestamp.
!F: convert a time duration in floating point seconds to Flux Standard Duration (FSD) string (e.g. {runtime!F}). Defaults to empty string if field does not exist.
!H: convert a time duration in floating point seconds to hours:minutes:seconds form (e.g. {runtime!H}). Defaults to empty string if time duration field does not exist.
!P: convert a floating point number into a percentage fitting in 5 characters including the "%" character. E.g. 0.5 becomes "50%" 0.015 becomes 1.5%, and 0.0005 becomes 0.05% etc.

As a reminder to the reader, some shells will interpret the exclamation point (!) when using a conversion flag. The exclamation point may need to be escaped (\!).

Annotations can be retrieved via the annotations field name. Specific keys and sub-object keys can be retrieved separated by a period ("."). For example, if the scheduler has annotated the job with a reason pending status, it can be retrieved via "{annotations.sched.reason_pending}".

As a convenience, the field names sched and user can be used as substitutions for annotations.sched and annotations.user. For example, a reason pending status can be retrieved via "{sched.reason_pending}".

The field names that can be specified are:

id: job ID
id.f58: job ID in RFC 19 F58 (base58) encoding
id.dec: job ID in decimal representation
id.hex: job ID in 0x prefix hexadecimal representation
id.dothex: job ID in dotted hexadecimal representation (xx.xx.xx.xx)
id.words: job ID in mnemonic encoding
id.emoji: job ID in emoji encoding
userid: job submitter's userid
username: job submitter's username
urgency: job urgency
priority: job priority
dependencies: list of any currently outstanding job dependencies
status: job status (DEPEND, SCHED, RUN, CLEANUP, COMPLETED, FAILED, CANCELED, or TIMEOUT)
status_abbrev: status but in a max 2 character abbreviation
status_abbrev: status but an appropriate emoji instead of job state / result
name: job name
cwd: job current working directory
queue: job queue
ntasks: job task count
ncores: job core count
duration: job duration in seconds
nnodes: job node count (if job ran / is running), empty string otherwise
ranks: job ranks (if job ran / is running), empty string otherwise
nodelist: job nodelist (if job ran / is running), empty string otherwise
state: job state (DEPEND, SCHED, RUN, CLEANUP, INACTIVE)
state_single: job state as a single character
state_emoji: job state but an appropriate emoji instead of DEPEND, SCHED, RUN, CLEANUP, or INACTIVE
result: job result if job is inactive (COMPLETED, FAILED, CANCELED, TIMEOUT), empty string otherwise
result_abbrev: result but in a max 2 character abbreviation
result_emoji: result but an appropriate emoji instead of COMPLETED, FAILED, CANCELED, or TIMEOUT
success: True of False if job completed successfully, empty string otherwise
waitstatus: The raw status of the job as returned by waitpid(2) if the job exited, otherwise an empty string. Note: waitstatus is the maximum wait status returned by all job shells in a job, which may not necessarily indicate the highest task wait status. (The job shell exits with the maximum task exit status, unless a task died due to a signal, in which case the shell exits with 128+signo)
returncode: The job return code if the job has exited, or an empty string if the job is still active. The return code of a job is the highest job shell exit code, or negative signal number if the job shell was terminated by a signal. If the job was canceled before it started, then the returncode is set to the special value -128.
exception.occurred: True of False if job had an exception, empty string otherwise
exception.severity: If exception.occurred True, the highest severity, empty string otherwise
exception.type: If exception.occurred True, the highest severity exception type, empty string otherwise
exception.note: If exception.occurred True, the highest severity exception note, empty string otherwise
t_submit: time job was submitted
t_depend: time job entered depend state
t_run: time job entered run state
t_cleanup: time job entered cleanup state
t_inactive: time job entered inactive state
runtime: job runtime
expiration: time at which job allocation was marked to expire
t_remaining: If job is running, amount of time remaining before expiration
annotations: annotations metadata, use "." to get specific keys
sched: short hand for annotations.sched
user: short hand for annotations.user

Field names which are specific to jobs which are also instances of Flux include:

instance.stats: a short string describing current job statistics for the instance of the form PD:{pending} R:{running} CD:{successful} F:{failed}
instance.stats.total: total number of jobs in any state in the instance.
instance.utilization: number of cores currently allocated divided by the total number of cores. Can be formatted as a percentage with !P, e.g. {instance.utilization!P:>4}.
instance.gpu_utilization: same as instance.utilization but for gpu resources
instance.progress: number of inactive jobs divided by the total number of jobs. Can be formatted as a percentage with {instance.progress!P:>4}
instance.resources.<state>.{ncores,ngpus}: number of cores, gpus in state state, where state can be all, up, down, allocated, or free, e.g. {instance.resources.all.ncores}

The following fields may return different information depending on the state of the job or other context:

contextual_info: Returns selected information based on the job's current state. If the job is in PRIORITY state, then the string priority-wait is returned, if the job is in DEPEND state, then a list of outstanding dependencies is returned, if the job is in SCHED state then an estimated time the job will run is returned (if the scheduler supports it). Otherwise, the assigned nodelist is returned (if resources were assigned).
contextual_info: Returns the job runtime for jobs in RUN state or later, otherwise the job duration (if set) is returned.
inactive_reason: If the job is inactive, returns the reason that the job is no longer active. Generally speaking, will output "Exit", "Timeout", "Canceled", or signal. If available, other contextual information will also be provided such as the exit returncode or cancellation message.

CONFIGURATION¶

The flux-jobs command supports registration of named output formats in configuration files. The command loads configuration files from flux-jobs.EXT from the following paths in order of increasing precedence:

$XDG_CONFIG_DIRS/flux or /etc/flux/xdg if XDG_CONFIG_DIRS is not set. Note that XDG_CONFIG_DIRS is traversed in reverse order such that entries first in the colon separated path are highest priority.

XDG_CONFIG_HOME/flux or $HOME/.config/flux if XDG_CONFIG_HOME is not set

where EXT can be one of toml, yaml, or json.

If there are multiple flux-jobs.* files found in a directory, then they are loaded in lexical order (i.e. .json first, then .toml, then .yaml)

Named formats are registered in a formats table or dictionary with a key per format pointing to a table or dictionary with the keys:

format: (required) The format string
description: (optional) A short description of the named format, displayed with flux jobs --format=help

If a format name is specified in more than one config file, then the last one loaded is used. Due to the order that flux-jobs loads config files, this allows user configuration to override system configuration. It is an error to override any internally defined formats (such as default).

If a format name or string is not specified on the command line the internally defined format default is used.

Example:

# $HOME/.config/flux/flux-jobs.toml

[formats.myformat]
description = "My useful format"
format = """\
{id.f58:>12} {name:>8.8} {t_submit!D:<19} \
{t_run!D:<19} {t_remaining!F}\
"""

It may be helpful to start with an existing named format by using the --format=get-config=NAME option, e.g.:

$ flux jobs --format=get-config=default >> ~/.config/flux/flux-jobs.toml

Be sure to change the name of the format string from default. It is an error to redefine the default format string.

EXAMPLES¶

The default output of flux-jobs(1) will list the pending and running jobs of the current user. It is equivalent to:

$ flux jobs --filter=pending,running

To list all pending, running, and inactive jobs, of the current user, you can use --filter option or the -a option:

$ flux jobs -a

OR

$ flux jobs --filter=pending,running,inactive

To alter which user's jobs are listed, specify the user with --user:

$ flux jobs --user=flux

Jobs that have finished may be filtered further by specifying if they have completed, failed, or were canceled. For example, the following will list the jobs that have failed or were canceled:

$ flux jobs --filter=failed,canceled

The --format option can be used to alter the output format or output additional information. For example, the following would output all jobids for the user in decimal form, and output any annotations the scheduler attached to each job:

$ flux jobs -a --format="{id} {annotations.sched}"

The following would output the job id and exception information, so a user can learn why a job failed.

$ flux jobs --filter=failed --format="{id} {exception.type} {exception.note}"

RESOURCES¶

Flux: http://flux-framework.org