Revision history for Tapper-MCP
5.0.0 2016-02-15
- internal refactoring
- use MooseX::Daemonize for daemon
- plugin system for organizing MCP functionality
- now as plugins: console handling, reset handling
- host pool handling: many "anonymous" hosts that have a
common name where only the count matters but not the
particular single name. Think of cloud instances.
- merged TestrunDB/ReportsDB
- polishing edge cases around:
* MCP chil exit handling
* state handling when installation is skipped
* output directory cleanup
* interdependent scenarios
* explicit db disconnects in MCP children
4.1.2 2012-11-23
- new precondition hint 'local' to run a PRC locally
around any program which is then your hook to any testing
(kind of a lightweight machine automation)
4.1.1 2012-11-20
- dzil pluginbundle TAPPER
4.1.0 2012-10-19
* Speedup rework (factor 5)
- [FEAT] Scheduler with non-DBIC objects
- [FEAT] scheduler in separate process
- [TEST] for memory growth
- make package immutable
* PostgreSQL support
- [FIX] slow first() usage
- rename USER into OWNER
* Cobbler integration (first part)
- [FEAT] new precondition_type "cobbler"
* Features
- [FEAT] host blacklists per queue ("denied host")
- [FEAT] poor mans nesting for SimNow
- [FEAT] new precondition "testrun_stop"
* Misc
- [FIX] condition to grep for forbidden hosts
- [FIX] job has only one queue
- [FIX] report correct testrun id to notification framework
- [FIX] add hostname feature if not in DB
- [FIX] wrong 'passwd' parameter name
- [FIX] deprecated "use UNIVERSAL"
- [RFK] adapt API to WFQ API
4.0.5 2012-06-26
- rework tmp/log/pid file handling in daemon and reset plugin
4.0.4 2012-06-12
- help with Crypt::DH deps via AAAA::Crypt::DH
4.0.3 2012-06-11
- hide test modules from PAUSE
4.0.2 2012-06-05
- timeout handling bit more lax to adapt PRC timing changes
- log files during test
4.0.1 2012-05-16
- *lots* of enhancements in whole automation layer
- MCP is now a stop/resume-able state engine,
based on event queues,
making the whole automation persistent,
inclusive still correctly recalculating timeouts
- introduce SSH-connect to test without complete machine setup
- introduce 'notification framework' to conditionally
trigger notifications on incoming results
- SIGUSR1 based live debugging, testrun-id part of process name
- introduced keep-alive detection
- reworked API for reset plugins
- OSRC-RESET plugin with watchdog-style success checking
- configurable and host-dependent grub-config
- improved integration with kickstart/autoyast/debinst installers
- allow lazy precondition producers also in "virt" preconditions
- big Dist::Zilla roundhouse kick
2011-03-30 Steffen Schwigon <steffen.schwigon@amd.com>
* v3.0.5
* rewritable shebang line
* yet another timeout increase
* [FIX] access to PRC number in message
* [FIX] handle warn-install message
* [FIX] provide testrun to TAP role
* [FIX] show error message when console can't be opened
* [FIX] provide MCP port
2011-03-10 Steffen Schwigon <steffen.schwigon@amd.com>
* v3.0.4
* distro maintenance
2011-03-08 Steffen Schwigon <steffen.schwigon@amd.com>
* v3.000002
* fix scheduler report sending with T::MCP::Net::TAP
* Make timouts more robust in tests
* try to force YAML versions
2011-03-02 Steffen Schwigon <steffen.schwigon@amd.com>
* v3.0.1
* cpan release