makeflow(1)

NAME

makeflow - workflow engine for executing distributed workflows

SYNOPSIS

makeflow [options] <dagfile>

DESCRIPTION

Makeflow is a workflow engine for distributed computing. It accepts a specification of a large amount of work to be performed, and runs it on remote machines in parallel where possible. In addition, Makeflow is fault-tolerant, so you can use it to coordinate very large tasks that may run for days or weeks in the face of failures. Makeflow is designed to be similar to Make, so if you can write a Makefile, then you can write a Makeflow.

You can run a Makeflow on your local machine to test it out. If you have a multi-core machine, then you can run multiple tasks simultaneously. If you have a Condor pool or a Sun Grid Engine batch system, then you can send your jobs there to run. If you don't already have a batch system, Makeflow comes with a system called Work Queue that will let you distribute the load across any collection of machines, large or small. Makeflow also supports execution in a Docker container, regardless of the batch system used.

OPTIONS

When makeflow is ran without arguments, it will attempt to execute the workflow specified by the Makeflow dagfile using the local execution engine.

Commands

-c,----clean <option>
Clean up: remove logfile and all targets. If option is one of [intermediates, outputs, cache], only indicated files are removed.
-f,--summary-log <file>
Write summary of workflow to file.
-h, --help Show this help screen.
-m,--email <email>
Email summary of workflow to address.
-v, --version Show version string.
-X,--chdir <directory>
Chdir to enable executing the Makefile in other directory.

Batch Options

-B,--batch-options <options>
Add these options to all batch submit files.
-j,--max-local <#>
Max number of local jobs to run at once. (default is # of cores)
-J,--max-remote <#>
Max number of remote jobs to run at once. (default is 1000 for -Twq, 100 otherwise)
-l,--makeflow-log <logfile>
Use this file for the makeflow log. (default is X.makeflowlog)
-L,--batch-log <logfile>
Use this file for the batch system log. (default is X.<type>log)
-R, --retry Automatically retry failed batch jobs up to 100 times.
-r,--retry-count <n>
Automatically retry failed batch jobs up to n times.
--wait-for-files-upto <#>
Wait for output files to be created upto this many seconds (e.g., to deal with NFS semantics).
-S,--submission-timeout <timeout>
Time to retry failed batch job submission. (default is 3600s)
-T,--batch-type <type>
Batch system type: local, dryrun, condor, sge, pbs, torque, blue_waters, slurm, moab, cluster, wq, amazon, mesos. (default is local)

Debugging Options

-d,--debug <subsystem>
Enable debugging for this subsystem.
-o,--debug-file <file>
Write debugging output to this file. By default, debugging is sent to stderr (":stderr"). You may specify logs be sent to stdout (":stdout"), to the system syslog (":syslog"), or to the systemd journal (":journal").
--verbose Display runtime progress on stdout.

WorkQueue Options

-a, --advertise Advertise the master information to a catalog server.
-C,--catalog-server <catalog>
Set catalog server to <catalog>. Format: HOSTNAME:PORT
-F,--wq-fast-abort <#>
WorkQueue fast abort multiplier. (default is deactivated)
-M,---N <project-name>
Set the project name to <project>.
-p,--port <port>
Port number to use with WorkQueue. (default is 9123, 0=arbitrary)
-Z,--port-file <file>
Select port at random and write it to this file. (default is disabled)
-P,--priority <integer>
Priority. Higher the value, higher the priority.
-W,--wq-schedule <mode>
WorkQueue scheduling algorithm. (time|files|fcfs)
-s,--password <pwfile>
Password file for authenticating workers.
--disable-cache Disable file caching (currently only Work Queue, default is false)
--work-queue-preferred-connection <connection>
Indicate preferred connection. Chose one of by_ip or by_hostname. (default is by_ip)

Monitor Options

--monitor <dir>
Enable the resource monitor, and write the monitor logs to
--monitor-with-time-series Enable monitor time series. (default is disabled)
--monitor-with-opened-files Enable monitoring of openened files. (default is disabled)
--monitor-interval <#>
Set monitor interval to <#> seconds. (default 1 second)
--monitor-log-fmt <fmt>
Format for monitor logs. (default resource-rule-%06.6d, %d -> rule number)
--allocation <waste>
When monitoring is enabled, automatically assign resource allocations to tasks. Makeflow will try to minimize waste or maximize throughput.

Umbrella Options

--umbrella-binary <filepath>
Umbrella binary for running every rule in a makeflow.
--umbrella-log-prefix <filepath>
Umbrella log file prefix for running every rule in a makeflow. (default is .umbrella.log)
--umbrella-mode <mode>
Umbrella execution mode for running every rule in a makeflow. (default is local)
--umbrella-spec <filepath>
Umbrella spec for running every rule in a makeflow.

Docker Support

--docker <image>
Run each task in the Docker container with this name. The image will be obtained via "docker pull" if it is not already available.
--docker-tar <tar>
Run each task in the Docker container given by this tar file. The image will be uploaded via "docker load" on each execution site.

Singularity Support

--singularity <image>
Run each task in the Singularity container with this name. The container will be created from the passed in image.

Amazon Options

--amazon-credentials <path>
Specify path to Amazon credentials file. The credentials file should be in the following JSON format:
{
"aws_access_key_id" : "AAABBBBCCCCDDD"
"aws_secret_access_key" : "AAABBBBCCCCDDDAAABBBBCCCCDDD"
}
--amazon-ami <image-id>
Specify an amazon machine image.

Mesos Options

--mesos-master <hostname>
Indicate the host name of preferred mesos master.
--mesos-path <filepath>
Indicate the path to mesos python2 site-packages.
--mesos-preload <library>
Indicate the linking libraries for running mesos..

Mountfile Support

--mounts <mountfile>
Use this file as a mountlist. Every line of a mountfile can be used to specify the source and target of each input dependency in the format of target source (Note there should be a space between target and source.).
--cache <cache_dir>
Use this dir as the cache for file dependencies.

Archiving Options

--archive <path>
Archive results of workflow at the specified path (by default /tmp/makeflow.archive.$UID) and use outputs of any archived jobs instead of re-executing job
--archive-read <path>
Only check to see if jobs have been cached and use outputs if it has been
--archive-write <path>
Write only results of each job to the archiving directory at the specified path

Other Options

-A, --disable-afs-check Disable the check for AFS. (experts only)
-z, --zero-length-error Force failure on zero-length output files.
--wrapper <script>
Wrap all commands with this script. Each rule's original recipe is appended to script or replaces the first occurrence of {} in script.
--wrapper-input <file>
Wrapper command requires this input file. This option may be specified more than once, defining an array of inputs. Additionally, each job executing a recipe has a unique integer identifier that replaces occurrences %% in file.
--wrapper-output <file>
Wrapper command requires this output file. This option may be specified more than once, defining an array of outputs. Additionally, each job executing a recipe has a unique integer identifier that replaces occurrences %% in file.
--enforcement Use Parrot to restrict access to the given inputs/outputs.
--parrot <path>
Path to parrot_run executable on the host system.
--shared-fs <dir>
Assume the given directory is a shared filesystem accessible at all execution sites.

DRYRUN MODE

When the batch system is set to -T <dryrun>, Makeflow runs as usual but does not actually execute jobs or modify the system. This is useful to check that wrappers and substitutions are applied as expected. In addition, Makeflow will write an equivalent shell script to the batch system log specified by -L <logfile>. This script will run the commands in serial that Makeflow would have run. This shell script format may be useful for archival purposes, since it does not depend on Makeflow.

ENVIRONMENT VARIABLES

The following environment variables will affect the execution of your Makeflow:

BATCH_OPTIONS

This corresponds to the -B <options> parameter and will pass extra batch options to the underlying execution engine.

MAKEFLOW_MAX_LOCAL_JOBS

This corresponds to the -j <#> parameter and will set the maximum number of local batch jobs. If a -j <#> parameter is specified, the minimum of the argument and the environment variable is used.

MAKEFLOW_MAX_REMOTE_JOBS

This corresponds to the -J <#> parameter and will set the maximum number of local batch jobs. If a -J <#> parameter is specified, the minimum of the argument and the environment variable is used.

Note that variables defined in your Makeflow are exported to the environment.

TCP_LOW_PORT

Inclusive low port in range used with -p 0.

TCP_HIGH_PORT

) Inclusive high port in range used with -p 0.

EXIT STATUS

On success, returns zero. On failure, returns non-zero.

EXAMPLES

Run makeflow locally with debugging:
makeflow -d all Makeflow
Run makeflow on Condor will special requirements:
makeflow -T condor -B "requirements = MachineGroup == 'ccl'" Makeflow
Run makeflow with WorkQueue using named workers:
makeflow -T wq -a -N project.name Makeflow
Create a directory containing all of the dependencies required to run the specified makeflow
makeflow -b bundle Makeflow

COPYRIGHT

The Cooperative Computing Tools are Copyright (C) 2003-2004 Douglas Thain and Copyright (C) 2005-2015 The University of Notre Dame. This software is distributed under the GNU General Public License. See the file COPYING for details.

SEE ALSO


CCTools 6.1.8 FINAL from source released on