Changing submitted job Settings, stalled nodes and quality of life suggestions

Render Manager Wishlist & Feature Requests
#1

Hi,

First post here. Love the the software and want to thank you for it first off. Scene manager in particular is something I have been hoping for for a long time now. I gave up on RenderPass Manager a few years ago now due to bugs and have been looking for some kind of alternative to managing scenes and different camera settings quickly and effectively especially on large scenes and animations. I also love how easily it integrates with your Render Manager which brings me to my point.

Having set up and run a few renderfarms at different studios with both backburner and deadline integrations there are a few things I am missing and I’m sure others are too.

  • Changing job settings after submission. This would be a massive time saver. Even just being able to change the rendered frame range and resolution would be huge and make all the difference. I’ve been there several times since i adopted the ecosystem where its late I want to get it rendering but have to reload the file to make these simple changes before resubmitting again. We’ve all been tired and forgot to set something up properly. This can get as granular as you like and becomes a bit of mission if you start catering to the different render engines so I would understand the reticence of going down that rabbit hole, but feel that frame range, and resolution should be something worth implementing for everyone.

  • Errored and stalled nodes. I’ve had this a few times, not sure exactly what’s causing it. That’s not the issue that’s my end. What I’m proposing is some kind of error limit or lives, where a stalled or failed node will get a certain amount of chances before its redlisted. So say it fails to renders or crashes out, rather than automatically failing that node allow the admin to set a tolerance so that it will try again rather than coming back to nothing rendered. Sometimes there is no reason for crashes that you can get to the bottom of and the node just needs to try again. A node restart node function could be something to explore further down the line, much like Thinkbox’s Pulse implementation but that a more complex animal.

  • Rerender failed tasks. This was a huge saviour in the past in deadline. Where nodes failed specific tasks in an animation sequence. You could simply right click and hit rerender failed tasks. It would then give you an option to black list any nodes that failed/errored out/are causing issues. This is especially helpful when its not a sequential block of frames but rather scattered throughout the sequence.

I love that you are trying to make renderfarm management easier and more straight forward. Coming from managing deadline renderfarms which is far and away more complicated than most would need for most arch viz pipelines and something that is more robust than backburner Pulze is a refreshing change and just works. If there was a way of getting these more in depth features in without breaking that ethos it would make a real difference in reducing time resending renders and just make job management that much easier and go a long way to breaking the hold that deadline has.

If these suggestions or issues have been raised before i can only apologise. I didn’t find anything in the forum and also couldn’t find anything in the documentation that would help with these specific issues.

Thanks

Curtis

#2

Dear Curtis,

Velcome in the community and thank you for your detailed suggestions and for your kind words!

First of all it is important to mention that all of your topics are under development and we are planning to release a major update in the beginning of the next year which will fulfill these needs for sure.

Please let me provide some additional info for your questions:

“Changing job settings after submission”

Yes, you are right it is an important feature and many users are notified that it is a missing one. So it will be implemented into the next version for sure.

In the meantime it is good to know that resending a job with the mentioned changes (frame range, resolution) is relatively easy. You do not need to open Max, you just have to browse the saved max file and you can set up these variables in the job submitter. As I mentioned the render settings are missing but it will be included soon.

“Errored and stalled nodes”

Yes, we have been working on the error handling heavily. We are experimenting with different approaches and we are looking for the most effective one. I just would like to highlight that at the moment you are able to restart RM on any PC remotely or restart any PC as an admin . Of course to schedule it and further automatize it will be the next step.

“Rerender failed tasks”

There are two ways at the moment to correct a failed frame:

If a specific Node fails a frame (renders it but not correctly) then you can open up the job, select a frame which is rendered by the faulty Node and hit “Select by Node” at the left side. It will select all the frames which are rendered by the faulty Node. After that you just have to “Reset” the selected tasks and other Nodes will render them correctly.

If there are frames which are missing from the output folder then you can use the “Missing frames” button. It will copy all the missing frame numbers onto the clipboard. Then create a new Job, browse the original max file, and paste the frame numbers into the “List” field. In this way just the Missing frames will be re-rendered.

We will further develop this feature for sure to make it more intuitive.

I’m very much pleased that you are seeing the big picture so consciously. Our goal is to provide a sophisticated and highly automated system under the hood but keep it simple and friendly for artists. We do our best and every feedback like yours supports us along the way. So thank you again!

Best regards,

Attila

selectbynode

1 Like
#3

Hi Attila,

Thank you for taking the time to get back to me with such a detailed response. Its great to hear that these things are already on your agenda, i had a feeling they would be anyway. Apologies if i maybe pointed out the obvious.

Also thank you for the instructions on how to work around my issues. I was aware of of the admin resets etc it was more atomization i was aiming at. Again great to hear you have this in hand. From my point of view a simple “cat has nine lives” approach is a good start for simple errored tasks. Ones where the node is still responding its just the task has failed to initiate or crashed out for whatever random IT gremlin its decided to find. Just to avoid the manually having to be log in and refresh from the node from the admin console. At least give the poor things a chance at automatic redemption before i have to step in and sort it. :slight_smile:

Your response here and elsewhere on the forum has only given me more confidence in what you are producing and I am 100% behind you guys. Your efforts are hugely appreciated and for such a young software solution to a VERY complex management problem I personally am very much behind the way you are taking it. I look forward to seeing how you iterate on your promising start.

Back to the more pressing problem I have, always need more nodes. Or at least another super powerful one :smiley: