CGRU Version 3.3.1 Released.

https://cgru.readthedocs.io/en/latest/changes_log/changes_log.html#v3-3-1
Various minor improvements.

CGRU Version 3.3.0 Released.

https://cgru.readthedocs.io/en/latest/changes_log/changes_log.html#v3-3-0

GPU resources monitoring.
CPU and GPU job block needs parameters.
Tasks depend mask cause Wait Dependencies new task state.

CGRU Version 3.2.2 Released.

https://cgru.readthedocs.io/en/latest/changes_log/changes_log.html#v3-2-2

Just lots of minor bug-fixes and enhancements accumulated.

CGRU Version 3.2.1 Released.

https://cgru.readthedocs.io/en/latest/changes_log/changes_log.html#v3-2-1

• Houdini Afanasy TOP Scheduler:
  https://cgru.readthedocs.io/en/latest/software/houdini.html#afanasy-top-scheduler
• Several AfWatch enhancements and bug-fixes.

CGRU Version 3.2.0 Released.

• Some configuration variables moved to pool parameters.
  So now it is more easy to setup each pool individually.
  It is especially useful, if your farm has a remote pool(s).
• Parser can return information about task process resources, server stores it, GUIs can show.
• RENDER_NO_TASK and RENDER_OVERLOAD events added.
• Server HTTP configuration extended.
  Now it is more easy to make server to serve some custom or even several WebGUIs.
• Several AfWatch bug-fixes.

CGRU Version 3.1.1 Released.

• AfWatch job item can collapse blocks.
  Useful for UI space economy, especially on jobs with a big blocks count.
  In a View Options menu you can collapse/expand all jobs,
  and set an option to collapse new jobs.
• Houdini submission ROP works with Python 3.
  Now you can choose to download Houdini with internal Python 3.

CGRU Version 3.1.0 Released.

Afanasy pool tickets got maximum hosts property.

This is mostly needed for licence hosts limits.
There is a common type of licensing where you can run multiple instances of software on same host, occupying only single license.

CGRU Version 3.0.0 Released.

Pools and Tickets

You can set root pool NUKE:20 pool tickets to limit Nuke licenses on the entire farm.
You can set some pool MEM:64 host tickets to limit RAM.
And set corresponding tickets to your job blocks.
So you will limil 20 NUKE tasks on the entire farm.
And each host can run 1 task with 64 MEM tickets, 2 tasks with 32, or 132 and 216 on each host.

Changes log:

https://cgru.readthedocs.io/en/latest/changes_log/changes_log.html#v3-0-0

Milestone:

https://github.com/CGRU/cgru/milestone/7?closed=1

CGRU Version 2.3.1 Released.

http://cgru.info/changes_log#release231

• Linux packages moved on SystemD.
• Windows services.
• NVidia GPU monitoring Python custom resource class.
• Bug fixes.

This is a minor release, the main reason is bug fixes.

CGRU Version 2.3.0 Released.

http://cgru.info/changes_log#release230

• Generally new Branch class appeared: http://cgru.info/afanasy/branch.
  Now jobs solving is going within branches hierarchy.
  Branches can represent departments/projects/assets and you can vary their priorities.
• Task progress change timeout - job block parameter.
• Task minimum run time - job block parameter.
• Numeric pattern frame number replacement is processed in a Python service class.
• Bug fixes.

This is the anniversary Afanasy release!
10 years ago it started to render frames in a real production: https://www.imdb.com/title/tt0493464/
Many major changes happened over that time, Afanasy became much better.
Lets remember the most significant architecture changes:

• Qt is used for GUI only. At first Afanasy was entirely written on Qt, so we needed to install lots of huge GUI libraries on each render node.
• JSON protocol. Very easy to communicate Afanasy. We do not need to compile af module for all needed Python versions any more.
• Web GUI. Browser gives us an ability to check farm without installing anything, we can do it just from a mobile phone.
• Server does not connect to clients. This was the first bottleneck for a bigger farms. When server connects to a big amount of clients it slows down. Now only clients connect to server.
• Server does not raise a thread for an each incoming connection. And this was the second bottleneck on a way to a bigger farms. Thread raising consumes much resources. Better to use threads pool (raise all needed threads once). Also you can configure server to use non-blocking IO (based on epoll facility so this is Linux only feature).
• Branches. You can tune farm resources allocation between departments/projects/assets.

This changes affected the project very much. I even want to say, that it is a completely another product than 10 years ago.
And Afanasy spreads more widely. From a few small studios with a local small farm to a bigger studios all over the world with hundreds of machines in a farm and in a cloud.
Many developers, TDs, render-wrangles, supervisors from different countries/studios contribute to this project.
Some of them write many things manually, some giving invaluable advises from their great experience, some even shared their in-house mature classes.
For a last 10 years working on the project we made it much-much better!
Many thanks for all this to all!