The iRODS Consortium and RENCI are pleased to announce iRODS 4.1.12.
This is mainly a bugfix release which addresses various replication and iscan/ifsck issues and provides additional debugging and visibility tooling for administrators.
This release will be the last in the 4.1 series. The 4-1-stable branch is now EOL.
CentOS6 packages are no longer supported (Our test framework requires Python 2.7+).
This release consists of 39 commits from 3 contributors and closed 36 issues.
The 4.1.12 packages are available at https://files.renci.org/pub/irods/releases/4.1.12/.
** Enhancements **
Add ichksum admin mode [#3265]
Add retry capability to replication resource [#3436] [#3746]
Add visibility for active rebalance operations [#3683]
Add enhanced_logging=1 option to server_config.json (4.1-only) [#3927]
** Bug Fixes **
Fix iscan exit code [#3256]
Marked as resolved/invalid [#3317] [#3719] [#3773] [#3789] [#3832] [#4106] [#4143]
Fix iscan for data in resource hierarchies [#3418] [#3419]
Fix ifsck for data in resource hierarchies [#3503] [#4107]
Fix rebalance operation to avoid data arriving after invocation [#3665]
Fix iscan to report missing zero-length files [#3681]
Fix MySQL case errors in group queries [#3717]
Marked as wontfix [#3759]
Fix uninitialized variables in datetimef() rule [#3767]
Documentation for decommissioning a resource [#3821]
Add testing for bad/corrupt file systems [#3854]
Fix database statement handling [#3862] [#4105]
Fix data placement logic with replication resources [#3904] [#3909]
Fix behavior in presence of root directory (/) used a vault path [#3928]
Fix icommands to report to ips correctly [#3991]
Fix iscan to page correctly when missing files detected [#4029]
Fix to only trigger replication when data changes [#4085] [#4099] [#4110]
Fix zone detection for federated clients [#4100]