This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
muster:9.0:release_notes [2019/04/23 14:55]
admin [9.0.13]
muster:9.0:release_notes [2019/08/30 10:15] (current)
Line 1: Line 1:
 ====== Release Notes ====== ====== Release Notes ======
 +===== 9.0.14 =====
 +  * [GENERAL] This version is LTS (Long Term Support). Future developments will be on Muster 9.5 while bug fixes will be carried on 9.0.14+
 +  * [GENERAL] Support for Blackmagic Fusion Command
 +  * [GENERAL] Support for Redshift standalone
 +  * [GENERAL] Support for reserve templates. They let you lock a render client by launching another application. This works well to integrate native render farm manager. Initial support includes support for Lux Core Slave, VRED Server Node, VRED Cluster Servce, CoronaDR, Lux Slave, VRay Spawner and VRaySwarm.
 +  * [GENERAL] A new dark VFX skin is enabled by default. Special thanks to Massive | A Ubisoft studio for it.
 +  * [GENERAL] Fixed Houdini PDG connector to handle cross platform scenarious and correctly handle backslashes on Windows supporting either UNC paths and drive mappings. Also code for distributed simulation has been included
 +  * [GENERAL] Released a new universal connector with initial support for Houdini. The Universal connector will incorporate under an unique code base connectors for softwares running Python 2/3 and Pyside/​Pyside2 (Maya, VRED...), it will be implemented in future versions
 +  * [GENERAL] Alarms handling has been moved on the Renderclient,​ they will handle them depending on the configuration and it is possible to write callbacks into the Global python template
 +  * [GENERAL] On multi instances setup, it's now possible for a job to borrow instances, by locking additional ones and freeing resources for intensive jobs. On GPUs engines like Houdini Redshift or Maya Redshift, GPUs allocated for the borrowed instances are automatically assigned to the instance holding the lock
 +  * [GENERAL] A new GPU affinity mask is available that works with Houdini Redshift/​Octane,​ and Maya Redshift/​Octane
 +  * [GENERAL] The history table now also holds the chunks owners
 +  * [GENERAL] Fixed an issue with maximum used instances when the jobs where not inside a folder
 +  * [GENERAL] OpenEXR supports now includes cropped images
 +  * [DISPATCHER] Database exports now correctly sets the IDs of the tables
 +  * [DISPATCHER] Exports of statistics in JSON format is now correctly encoded
 +  * [DISPATCHER] Fixed a bug where a keep alive message may be sent to pending invalid connections crashing the Dispatcher
 +  * [DISPATCHER] Fixed an error in the setDelay() function not working as intended since several versions
 +  * [DISPATCHER] Fixed an issue in the SQL Server driver where the db_jobs_data were not indexed correctly
 +  * [DISPATCHER] Job names can have more than 100 characters, previously it caused a crash in the log filename generation
 +  * [DISPATCHER] It's now possible to limit a job to a minimum amount of physical GPUs
 +  * [DISPATCHER] A review of the connected accounts is available in the Dispatcher statistics view
 +  * [DISPATCHER] Timed paused and resume on instances are now available
 +  * [DISPATCHER] Timed paused and resume on jobs have been fixed, they were not working under certain circumstances ​
 +  * [DISPATCHER] A license count check is now available on a job basis, this allows to check for a dual licensing (like Houdini and Redshift that have different licenses servers)
 +  * [DISPATCHER] Fixed an issue with SQL Server not able to store db_nodes data complaining shared_log_paths value was null
 +  * [DISPATCHER] Fixed wakeup at farm full load and required by pool functions that were considering folders too triggering the start even when it was not necessary
 +  * [CONSOLE] New filters available in Console and web interface for templates and pools
 +  * [CONSOLE] Is now possible to copy/paste user's and group'​s rights
 +  * [CONSOLE] Host views and pools manager now reports the instance GPUs
 +  * [CONSOLE] Is now not possible to remove or rename the currently logged user from Console
 +  * [CONSOLE] Fixed an error in the script editor where the windows disappear if the ESC key is pressed
 +  * [CONSOLE] Python console and code editor have the font size configurable
 +  * [CONSOLE] If an instance has CPU/GPU affinity enabled, it is shown in a dedicated column
 +  * [WEB] Configurations from the WEB REST APIs can be read only by Administrator,​ previously they were exposed to any user
 +  * [WEB] A new WEB API /​api/​profile returns features and capabilities of the currently logged user
 +  * [WEB] Is now possible to browse a chunk'​s log from the web interface
 +  * [WINDOWS] The entire code base of Windows has been changed to use only Unicode functions. This incremented the overall compatibility on the Windows platform as well as handling Unicode natively
 +  * [LINUX] Fixed an error in Python startup where in daemon mode the stdout/​stderr were closed and this caused startup errors in the systemd scripts under certain circumnstances
 +  * [LINUX] Fixed a crash in Linux happening when sending a command line with more than 255 arguments (yes that happens!)
 +  * [LINUX] Fixed an issue in the Houdini IFD template on the Linux platform not passing the correct arguments to the executables
 +  * [API] the MInstanceAPI now has a logAddMessage to print a log entry from Python
 +  * [API] MInstanceAPI.getRunningJob() and MInstanceAPI.getRunningChunk() correctly returns None if no job is running
 +  * [API] Add ability to reset the chunks require from the MDispatcherAPI and MConsoleAPI modules
 +  ​
 ===== 9.0.13 ===== ===== 9.0.13 =====