[Tok_Tok]

Hello,

I would really like to see network rendering be a bit more stable, and by "a bit" I mean a lot! The Manager often crashes for no reason and when it crashes it takes the nodes with it. Next to that you lose all of the jobs you put in the queue

This is a really big thing for me because no matter how much I would like to see speed improvements or GPU support, when your renders crash in the middle of a job and find out hours later, you are loosing a LOT of time.

Maxwell is an excellent program in many ways, but network stability is not one of them. This is a big wish for me.

[T0M0]

^THIS!

My personal wishlist (sorted by priority):
1. stable Maxwell Render network - XML errors
2. better and responsive forum support - some threads are with very late answer from dev's or even without answer
3. GPU acceleration for production engine - yeah, would be great

[ababak]

+1 for rock-solid network rendering.

Here are some ideas:
1) Render nodes should be able to continue rendering even if the connection with the manager is lost.
2) Render nodes should by all means be able to recover after crash and either continue rendering or start a new render with different processor ID to be able to send several mxi files to merge. Ideally they should be able to recover even while the manager is not available.
3) Render manager should be able to recover after crash. It should start listening to the network to understand what's going on and be able to take over all the running jobs.
4) Render manager should be able to talk to reconnected render node and download what has been done before the disconnection occurred.
5) Render manager should not rely on file systems and paths. It should store the results locally (or on the mounted network volume)

Comments and suggestions are welcome.

[Tok_Tok]

+1 for rock-solid network rendering.

Here are some ideas:
1) Render nodes should be able to continue rendering even if the connection with the manager is lost.
2) Render nodes should by all means be able to recover after crash and either continue rendering or start a new render with different processor ID to be able to send several mxi files to merge. Ideally they should be able to recover even while the manager is not available.
3) Render manager should be able to recover after crash. It should start listening to the network to understand what's going on and be able to take over all the running jobs.
4) Render manager should be able to talk to reconnected render node and download what has been done before the disconnection occurred.
5) Render manager should not rely on file systems and paths. It should store the results locally (or on the mounted network volume)

Comments and suggestions are welcome.

Excellent list! I would like to add that it should be possible to sneeze in your own home without it crashing.. I just had another crash, hours of rendering wasted

[astanelle]

1+ for a more healthy rendernetwork.

Unfortunately there are little infos / news / response to the NetworkTP.
And for a MAC user it's still kind of unusable.(Reading MXS issue).

Beside that i would like to see a more responsive forum as well. Maybe a clean up / restart would be helpful too.

[Nova66]

What more can I say except +1

Andrew.