Project

General

Profile

Actions

Bug #5426

closed

workflow lsid changes when running from command line

Added by Daniel Crawl almost 13 years ago. Updated over 11 years ago.

Status:
Resolved
Priority:
Normal
Assignee:
Category:
core
Target version:
Start date:
06/20/2011
Due date:
% Done:

0%

Estimated time:
Bugzilla-Id:
5426

Description

There are two problems that occur when running a workflow from the command line and specifying a non-default parameter value.

The first is that the revision is always incremented if the parameter value is the same:

./kepler.sh -runwf -nogui -x 4 ~/KeplerData/paramx.xml

WF LSID is urn:lsid:kepler-project.org/ns/:6018:25:2

./kepler.sh -runwf -nogui -x 4 ~/KeplerData/paramx.xml

WF LSID is urn:lsid:kepler-project.org/ns/:6018:25:3

The second is that if .kepler and KeplerData/modules/core/coreDB do not exist, a new LSID is used:

./kepler.sh -runwf -nogui -x 4 ~/KeplerData/paramx.xml

WF LSID is urn:lsid:kepler-project.org/ns/:6020:1:1

For desktop use, .kepler and KeplerData usually do not get deleted, but when executing on a cluster, these directories are not always available. This causes problems for provenance, e.g., the same workflow run on 10 different nodes will be identified as 10 different workflows instead of 10 runs of the same workflow.


Files

paramx.xml (23.6 KB) paramx.xml Daniel Crawl, 06/20/2011 11:20 AM
Actions #1

Updated by Daniel Crawl almost 13 years ago

Attaching example workflow.

Actions #2

Updated by Derik Barseghian over 12 years ago

Seems like an advanced user usecase, so retargeting post reporting-2.3.0. Please put it back if you feel otherwise.

Actions #3

Updated by Daniel Crawl almost 12 years ago

fixed: kepler.sh -runwf -noilwc ... will not change the lsid.

Actions #4

Updated by jianwu jianwu almost 12 years ago

I re-open this bug since we agreed that we need a versioning model for Kepler during Kepler 2.4 release conference call at June 15 2012.

[6/15/12 10:42:45 AM] Ilkay Altintas: We need to discuss what is a workflow change and come up with a new versioning model
[6/15/12 10:43:00 AM] Ilkay Altintas: (Related to Bug 5426)
[6/15/12 10:44:11 AM] Ilkay Altintas: Christopher mentioned a new (passoro?) system for high-energy physics that does versioning for actors
[6/15/12 10:45:07 AM] Christopher Brooks: triquetrum.org Feel free to request membership in the CHESS triq workgroup. The mailing list does not have much traffic.
[6/15/12 10:50:00 AM] Ilkay Altintas: We should check with Triq if there's anything that we could reuse.

Actions #5

Updated by Ilkay Altintas over 11 years ago

Added a command line option to stop it. Closed for 2.4 release. For 2.5, add a new bug with investigation of alternatives. We need to make sure this links to the new bug.

Actions #6

Updated by Daniel Crawl over 11 years ago

I created #5687 for 2.5.

Actions #7

Updated by Redmine Admin about 11 years ago

Original Bugzilla ID was 5426

Actions

Also available in: Atom PDF