1. Pegasus 5.1.x Series

1.1. Pegasus 5.1.2

Release Date: Feb 3rd, 2026

We are happy to announce the release of Pegasus 5.1.2. It is a minor release in the 5.1 branch. We invite our users to give it a try.

The release can be downloaded from: https://pegasus.isi.edu/downloads

1.1.1. Highlights of the Release

  1. Move to Condor File IO for OSDF transfers.

    OSDF transfers are now always delagated to HTCondor to manage using HTCondor file IO, especially when turning on Bypass Input File Staging. This applicable both for condorio and nonsharedfs data configurations.

    More details can be found in the documentation.

  2. Support for Flux This release of Pegasus has support for running workflows on HPC resources managed by Flux resource manager. This support relies on changes to HTCondor which will be made available in an upcoming 25.7.0 release scheduled for March 2026.

    Details of mapping Pegasus resource profiles to flux parameters can be found here

    This repository has useful scripts that a user can use to deploy pegasus in user mode on Flux Systems.

  3. Modified Pegasus versioning scheme to use Semantic Versioning Scheme v2 #2126

  4. Support for Python 3.14

1.1.2. New Features and Improvements

  • avoid parsing sub workflows into memory when parsing the top level workflow that includes them #2148

  • flux support #2143

  • Move OSDF transfers to be via condor file transfers in Pegasus Lite instead of relying on pegasus-transfer #2141

  • pick user provided env script for PegasusLite from the site where the job runs #2136

  • limit the number of pegasus-monitord launches in pegasus-dagman #2134

  • Add support for Python 3.14 #2128

  • Modify Pegasus versioning scheme to use Semantic Versioning Scheme v2 #2126

  • enable condorio support for bosco/ssh style job submissions for remote HPC clusters #2121

  • use job classad variables to shorten paths in transfer_input_files key in the job submit directories #2120

  • Explore condor_dag_checker - something we want to use as part of planning? #2116

  • update sqlite jar to latest stable 3.49.1.0 or higher #2109

  • [PM-1833] pmc cpuinfo invalid detection #1946

  • [PM-1098] encrypt the credentials when transferred with jobs #1212

  • [PM-1092] Ask Condor team to propagate glite errors in gridmanager #1206

1.1.3. Bugs Fixed

  • enforce maximum document parsing size for yaml docs to 2047 MB #2152

  • cpu atttributes are not included in the job composite event #2150

  • ensure user JAVA_HEAPMAX and JAVA_HEAPMIN values are propagated to the planner invocations for sub workflows #2147

  • bypass in condorio mode gets incorrectly triggered if a directory path specified for a local file #2142

  • cleanup jobs running remotely in nonsharedfs get associated with a container #2137

  • CLI tools pollute PYTHONPATH #2135

  • condor quoting is not triggered for arguments for glite and ssh style jobs #2132

  • monitord fails to parse job.out file if location record is malformed #2131

  • pegasus-init remote cluster option creates incorrect paths for local site #2125

  • worker package staging broken in sharedfs for create dir job (if set to run remotely) #2124

  • monitord overwrites transfer_attempts records found in the job.out file #2123

  • pegasus-analyzer –debug-job option broken #2122

  • pegasus-mpi-cluster: cpuinfo validation fails on hybrid CPU architectures #2119

1.1.4. Merged Pull Requests

  • Bugfix: removes incorrect use of the ‘local’ keyword in flux_local_submit_attributes.sh #2145 (ilumsden)

  • Adds a “local_submit_attributes.sh” script for the Flux RJMS #2144 (ilumsden)

  • Update pegasus-halt to fix POSIX compliance issue, bug on OSs where SH-> dash #2139 (ahnitz)

1.2. Pegasus 5.1.1 and 5.1.0

Release Date: May 29, 2025

We are happy to announce the release of Pegasus 5.1.1. It is a minor release on top of Pegasus 5.1.0 which is a major release of Pegasus. It also includes all features and bug fixes from the 5.0 branch. We invite our users to give it a try.

We recommend that users upgrade to 5.1.1 and not 5.1.0 because 5.1.1 has a fix for #2113 .

The release can be downloaded from: https://pegasus.isi.edu/downloads

If you are an existing user, please carefully follow these instructions to upgrade at https://pegasus.isi.edu/docs/5.1.0/user-guide/migration.html#migrating-from-pegasus-5-0-x-to-pegasus-5-1

1.2.1. Highlights of the Release

  1. Refined Data Transfer Mechanisms for Containerized Jobs

    PegasusLite now offers two distinct approaches for handling data transfers in
    containerized jobs. The shift to host-based transfers as the default aims to simplify workflows and minimize the overhead associated with customizing container images.

    • Host-Based Transfers (Default in 5.1.0): Input and output data are staged on the host operating system before launching the container. This method utilizes pre-installed data transfer tools on the host, reducing the need for additional configurations within the container.

    • Container-Based Transfers: Data transfers occur within the container prior to executing user code. This approach requires the container image to include necessary data transfer utilities like curl, ftp, or globus-online. Users preferring this method can set the property pegasus.transfer.container.onhost to false in their configuration files.

    More details can be found in the documentation.

  2. Integration with HTCondor’s Container Universe

    Pegasus 5.1.0 introduces support for HTCondor’s container universe, which is useful in pure HTCondor environments such as PATh/OSPool whereby the container management is handled by HTCondor. This integration simplifies job submission and execution, for environments where HTCondor’s container universe is available.

    This enhancement builds upon Pegasus’s initial container support introduced in version 4.8.0, reflecting ongoing efforts to improve compatibility and user experience.

    More details can be found in the documentation.

  3. pegasus-status command line tool was rewritten in python, removing Pegasus perl dependency. The new pegasus-status command has better support for showing status of hierarchical workflows.

  4. Improved determination on what site a job runs on. Starting Pegasus 5.1.0 release, PegasusLite wrapped jobs send a location record that enables us to figure out what resource a job runs on. The location record can be found toward the end of the job .out file as a pegasus multipart record.

    More details can be found in the documentation.

  5. Please note that RPM packaging for 5.1.x series is not compatible with the 5.0.x series. If you try to update an existing 5.0.x install you will see an error similar to the trace below

dnf update pegasus
...
Running transaction check
Transaction check succeeded.
Running transaction test
The downloaded packages were saved in cache until the next successful transaction.
You can remove cached packages by executing 'dnf clean packages'.
Error: Transaction test error:
 file /usr/lib64/pegasus/python from install of pegasus-5.1.0-1.el8.x86_64 conflicts with file from package pegasus-5.0.9-1.el8.x86_64

The recommended way is to first remove the 5.0.x install and then do the install.

1.2.2. New Features and Improvements

  1. update deployment scenarios documentation to include Open OnDemand configuration #2112

  2. Incorporate release notes into the documentation #2111

  3. document use of containers on HPC clusters #2110

  4. update planner worker package staging logic to default to rhel8 for linux and macos_14 for macos #2108

  5. planner should try and visualize the workflow #2099

  6. CVE-2025-21502 in RHEL 8 project dependency java-11-openjdk #2097

  7. Failures when testing pegasus v5.1.0 release #2095

  8. Improved determination of what resource the job ran on #2094

  9. add support in pegasus-init for submission to remote SLURM cluster via SSH #2093

  10. [PM-1999] pick automatically system executables on the local site on the basis of what PATH is set when running the planner #2091

  11. [PM-1975] enable bypass staging of container when running in container universe #2081

  12. [PM-1971] Update jars that have security vulnerabilities. #2077

  13. [PM-1970] Storage constraint test failing #2076

  14. [PM-1969] Schema doc should be for YAML and not XML #2075

  15. [PM-1956] p-version timestamp issue #2069

  16. [PM-1950] enable users to use container universe when running containerized jobs in pure condor environments #2063

  17. [PM-1942] Support transfers for a job on the HOST OS instead of from within the container #2055

  18. [PM-1901] converting pegasus-analyzer tool to API (like status) and add test suite #2014

  19. [PM-1889] Escape command line args passed to Job #2002

  20. [PM-1886] Recommendations for new Pegasus-Status CLI tool #1999

  21. [PM-1882] replace the perl command line client with python client #1995

  22. [PM-1881] remove dependency on pegasus-status command line tool , in the Workflow status function #1994

  23. [PM-1775] add changes to the Python API to support adding of checkpoint files by pattern #1889

  24. [PM-1589] Example workflows repository for 5.0 #1703

  25. [PM-1974] update InPlace cleanup algorithm to delete container image from the user submit directory #2080

  26. [PM-1967] Kickstart should kill a job gracefully before maxwalltime #2073

  27. [PM-1955] Deprecate R API #2068

  28. [PM-1934] source builds with multiple python3 installs #2047

  29. [PM-1915] Convert pegasus-statistics tool to API and add test suite #2028

  30. [PM-1914] update python workflow api to support arm64 #2027

  31. [PM-1912] planner should keep in mind units when converting diskspace profiles #2025

  32. [PM-1860] aws batch support needs to pick up credentials.conf correctly #1973

  33. [PM-1819] 5.0.3 Python API Improvements #1932

  34. [PM-1801] sqlalchemy warnings against 5.0 database #1915

  35. [PM-1793] refactor pegasus-transfer so that it can be invoked directly from pegasus-checkpoint #1907

  36. [PM-1782] incorporate pegasus arm builds into our build infrastructure #1896

  37. [PM-1781] pegasus-keg sleep option #1895

  38. [PM-1756] paths with spaces need to be escaped #1870

  39. [PM-1690] the –json option added in pegasus-plan/run needs to be integrated into the python api client code #1804

1.2.3. Bugs Fixed

  1. snakeyaml version version 1.32, used by Jackson 2.14 has an in-built circuit breaker that breaks parsing for large yaml documents #2113

  2. pegasus aws batch test failing because of urllib3 incompatibility #2107

  3. Planner should catch deep lfn common name problem when using CEDAR #2106

  4. support for condorio deep LFN broke after move to host OS based transfers #2105

  5. when parsing container mount points in the TC, normalize the path to ensure any duplicate / in directory paths are removed #2103

  6. pegasus-graphviz fails on a wf generated with java dax api that has no jobs #2101

  7. [PM-1954] Importing six.moves raises ModuleNotFoundError on Python 3.12 #2067

  8. [PM-1968] pegasus.gridstart allows values that are not documented #2074

  9. [PM-1952] Local universe job fail with pegasus.transfer.bypass.input.staging = true #2065

  10. [PM-1924] API submit still drops debug info to stdout #2037

  11. [PM-1902] Pika problem on RHEL 9 - Bump the version to 1.2.1 also #2015

  12. [PM-1923] download form does not send metrics to metric server #2036

1.2.4. Merged pull requests