Max freezes with distributed render

[Otuama]

Hi.

I've been using Corona 3 for a few months.... No problems with it.

Everyone else in the office updated to it yesterday because I was on a tight deadline and needed to use distributed.

It renders fine.  When finishing the box comes up saying it's collecting information.

When the box goes, it doesn't do anything.  Max just hangs.  It doesn't get to the denoising stage.

After a while of waiting..... hoping it'll wake up, I had to crash max.

This happened a couple of times.

I had to load the scene on a faster system and render locally.

Still missed my dealine :( .

Anyone had the same problem?

[aaouviz]

I think I'm experiencing the same issue.

I'm currently rendering out 7k shots (for the first time, so I originally thought it was simply a result of the slave not handling the workload), and overnight every time Max has frozen up towards the end of the render. As I haven't been watching the screen at the time, I can't say exactly a which point it's happening.

However, I have found a possible work-around. This morning I had the same problem (woke up to a frozen max/corona)... I simply turned off the slave PC and the master seemed to get unstuck.

Try this next time, hope it helps...

[maru]

We'd love to help, but we will definitely need more info:

1) Is this ALWAYS happening, or only sometimes, or randomly?
2) Is this happening in ANY scene (even super simple like just one teapot), or only in one/some specific scenes?
3) When Max is frozen on the master machine, please capture a minidump and send it to us - see point 2 here: https://coronarenderer.freshdesk.com/support/solutions/articles/5000524006
4) When Max is froze on the master machine, go to the slave machine, and capture a minidump from drserver.exe and 3dsmax.exe (on the slave) - see the above guide

You can host your files here: https://corona-renderer.com/upload
We can continue here, but I would definitely advise to start a support ticket about your issue: https://coronarenderer.freshdesk.com/support/tickets/new

Also, please make sure that you have the same 3 hotfix 1 version of Corona installed on all computers, and that you are running DR server version 3 hotfix 1 on all slave PCs.

[aaouviz]

Hi Maru,

I'm neck deep in a tight deadline at the moment, so I'd love to help and debug this but right now I can't unfortunately.

To answer your questions as best as I can:
- As far as I can tell, this is always happening. Though it's only happening when I use 3dsmax batch render, and it's only for the last of the batch. Oddly enough.
- It's happening on multiple scenes, although they are fairly heavy scenes.
- Ok, I should be able to do this. Though it's not exactly freezing, so much as it is 'hanging' until I turn off the slave.
- Will do.

Both PC's have the same v3 hotfix 1 installed.

Thanks for the reply - will try to get more info to you soon.

[Dionysios.TS]

Hi Maru,

I'm neck deep in a tight deadline at the moment, so I'd love to help and debug this but right now I can't unfortunately.

To answer your questions as best as I can:
- As far as I can tell, this is always happening. Though it's only happening when I use 3dsmax batch render, and it's only for the last of the batch. Oddly enough.
- It's happening on multiple scenes, although they are fairly heavy scenes.
- Ok, I should be able to do this. Though it's not exactly freezing, so much as it is 'hanging' until I turn off the slave.
- Will do.

Both PC's have the same v3 hotfix 1 installed.

Thanks for the reply - will try to get more info to you soon.

We still having freezing problems here and there. Never found out the reason. Only thing I know is that randomly one DR Server at a time freezes the process.
Not very happy about it. I think the Corona team should send for a day or two somebody to look at it in first person and understand what's going on!
Second solution, not in process yet, change render engine, something I would like to avoid!

[NormAndTheGang]

Hey guys,

We've been having the same issue for about a month now.

When we render without distributed rendering there isn't any problem but when we render locally with distributed rendering or through deadline with DBR have this problem more and more often. I'd say it's about 20-25% of our render that are having this issue.
It's becoming a real problem as with tight timelines we can't afford to have an image render for 4-5 hours and not be able to use it one time out of 4.

I just saw the minidump message so next time it happens I'll send it your way.

Chris

[TomG]

Also, are you using bloom and glare in these scenes? If it is freezing only at the end, I have noticed that when denoising is running AND the bloom and glare status bar is updating, it can cause a hang (as part of the known and tracked issue about bloom and glare progress bar hanging unless you swap to be looking at the Stats tab in the VFB rather than the post process)

[aaouviz]

Hi,

This is still an issue for me. I'm neck deep in deadlines, so still unable to de-bug it, but it's certainly causing headaches.

The previous comment seems true (without proof on my end) that it happens during denoising/bloom+glare ... and when doing Batch render with DR.

As soon as I close the DR on the slave, the render on the master PC finalises and emerges from it's hung state.

I'm using Corona 3, Hotfix 1 on both machines. (Will update to daily shortly - and test to see if it's still happening.)

[Dionysios.TS]

We have many cases like this. Is really very difficult to deal with, we also have a lot of deadlines and find a list of jobs not done cause of the freezing problems is very frustrating.

Personally, I think this is a high priority issue, more than having new features in the engine. What we can do with the new features if the engine freezes all time?

I started mentioning this problem since September 2018. Now we are in February 2019 and we have no clues why is happening. Very sad...

[maru]

Everyone with this issue, please see this post: https://forum.corona-renderer.com/index.php?topic=23266.msg142184#msg142184
Collect all the information, and send it to us using https://coronarenderer.freshdesk.com/support/tickets/new
We can't promise we will immediately solve it, but more reports mean more details and higher chance of having it fixed.

I can perfectly understand your frustration, but just reporting in on the forum and confirming it's not working for you is not going to help much.

The biggest issue with DR issues is that generally it is intended to be a "one click" solution. You just start DR server on the node(s), tick DR on the master, start rendering, and it just works. It works like this for most users, but some other ones are having all kinds of issues, which are extremely hard to diagnose.
The most common causes of issues are:
-Antivirus/firewall - this includes Windows Firewall and Defender - the easiest way to diagnose this is to completely disable them, and try DR again. If it works, then you should troubleshoot further, for example by adding rules to Windows Firewall - https://coronarenderer.freshdesk.com/support/solutions/articles/12000050816 - this is known to help a lot of people with similar problems!
-Another most common cause is some kind of network glitch - this is ultra hard to diagnose - for example, sometimes issues disappear when the router is restarted, or when its settings are reverted to defaults.