Skip to content

workflow.status file

Chris Churas edited this page Mar 20, 2015 · 20 revisions

NOTE: Service is currently NOT parsing this output, but there is a plan to do it (https://support.crbs.ucsd.edu/browse/CS-137)

The workflow.status file lets the workflow developer inform the service of the current status of the workflow.

Put the file in $CWS_outputdir directory

This file follows the java properties file format:

KEY=VALUE

The workflow developer can set values for the following KEYS

  • time

  • Current time of update in seconds since epoch since 1969

  • estimated.total.cpu.seconds & ..help

  • If known estimated amount of CPU in seconds this job will consume. This is different then the walltime. It is the the total amount of compute this workflow will consume. The estimated.total.cpu.seconds.help field lets caller set information about the estimate that can be displayed to the user.

  • estimated.total.diskspace & ..help

  • If known, estimated amount of disk in bytes this job will consume. The estimated.total.diskspace.help field lets caller set information about the estimate that can be displayed to the user.

Example:

#Time this file was updated
time = # unix time of event ie seconds since 1969 aka `date +%s`

# This is the wall time the workflow will take to run
estimated.walltime.seconds = # estimated walltime of job
estimated.walltime.seconds.help =

# This is the estimated total cpu time the workflow will take to run
estimated.total.cpu.seconds = # seconds of cpu time this workflow will consume
estimated.total.cpu.seconds.help =

# Estimated total disk space for job.  
estimated.total.diskspace =
estimated.total.diskspace.help =

# Denotes current phase of processing, should be a phase in phase.list below
phase =
phase.help =

# Lists the phases that this workflow runs through
phase.list =
phase.list.help =

#
# the format of this field is clustername:seconds,clustername:seconds
# If clustername is not known just use 'unknown'
#
cpu.seconds.consumed.per.cluster.list=
cpu.seconds.consumed.per.cluster.list.help=

# actual disk space consumed by workflow
diskspace.consumed =
diskspace.consumed.help =