Hello reader,
It's time for another development update. For the server, the trim API received an important bug fix so that it can no longer trim the last replica. The --link
option for iput
and irsync
was updated to correctly ignore symlinks. Before this change, the --link
option would result in immediate termination of the operation.
There's been a huge amount of activity in regard to the build and packaging effort. We now offer packages built against libc++ (for the 4.3 series) and libstdc++ (for the upcoming 5.x series). Dependency declarations for externals packages are better. The Clang compiler has been bumped to version 13.0.1. Users experiencing packaging issues involving the MySQL database plugin and MariaDB packages will be happy to know we've resolved the problem. And finally, the CMake package version file has been updated to meet the requirements of the version numbering change. As a result, the Docker-based Development Environment has been updated to support libstdc++.
Work on the Kubernetes Testing Environment is progressing. For those unaware, we've been working on new test infrastructure designed to run all tests on every GitHub pull request, hands-free. As of now, we can launch iRODS Providers and Consumers and run core tests. There's even a GUI! We see a lot of promise, but there's still a lot of work to do.
We have released the Indexing capability plugin for iRODS 4.3.1. The plugin now uses Boost.Beast for HTTP communication with Elasticsearch. This release includes support for TLS communication and Basic authentication with Elasticsearch. The plugin is designed to work with Elasticsearch 7 and later. An important note about this release - the document-type rule engine plugin is no longer provided by the Indexing capability plugin and as a result, you'll need to remove the document-type rule engine plugin from your server_config.json file. See the README at https://github.com/irods/irods_capability_indexing/tree/88bbf7ba6a8276c1349ad41afb0346fbb0ca9ac8 for details.
Metadata Guard 4.3.1.1 is almost ready. The release will land shortly after this update. The big news for this release is it now understands the atomic metadata operations API PEPs.
The Python iRODS Client (PRC) is receiving steady improvements. Work to add support for automated testing via GitHub actions is slowly taking shape. Support for PAM is being improved. Specifically, the exception information will be preserved following PAM login failures. Last but not least, the PRC is growing iinit-like capabilities such as auto-refresh and initialization of PAM sessions.
We'll end things with this ... iRODS S3 API 0.2.0 is now available! This release includes 25 closed issues providing enhancements such as multipart upload, connection pooling, configuration validation, etc. To view the list of closed issues, see https://github.com/irods/irods_client_s3_api/milestone/2?closed=1. Docker images are available at Docker Hub.
See you next month!
Metadata Templates Working Group
New Development Work
Active Development Work
Build and Packaging
Docker-based Development Environment
iRODS Kubernetes Testing Environment
Background Items
Docker-based Testing Environment
Unified Storage Tiering Capability
iRODS Command Line Interface (CLI)
Policy Composition
Logical Locking
New RPC API framework
Parallel Filesystem Integration
CockroachDB Database Plugin
Discussion
rcDataObjTruncate - deprecation and new API endpoint
iRODS 5.0.0 - hostnames and ephemeral servers