watchmaker crash Down phenomenon after 25 minutes of watchout production

[Gates kweon]

Down phenomenon after 25 minutes of watchout production

After setting the watchout 10 days ago, it was played without any problem.
The sauce was changed once a day.

And 3 days ago, the abnormal symptoms started.

Immediately after 24 minutes of watchout production
will be shut down.

We haven't figured out the cause yet.

These are the actions I have taken.
1. Test by version 6.6.5~6.6.7
2. Test by codec (HAP,H264,PROLESS, ..MOV,MP4..)
3. Windows Security Update
4. Test this file in another production..(same symptom)
5. Tested this file on my laptop.. (This works fine)
6.NVIDA VERSION No problem..

Videos and files are attached below.
Seeking advice from people who have experienced these symptoms.

[RBeddig]

If I understand it correctly, the show crashes on the production computer when the timeline reaches 25+ minutes?

Have you checked the log files on the production computer?

Have you checked the event log files (application) in Windows?

Have you deleted the WATCHOUT cache file on the production computer so that it gets rebuilt?

Can you rule out that some media at that point in the timeline is corrupt?

[Gates kweon]

Thanks for the reply.

I only backed up the mapping data and worked again.

And the video file is completely new and only one file
I did a test.

Still had the same symptoms.
(Then I think log files and cache files are irrelevant.)

I think it's another bug, how do you see it?

[RBeddig]

Well, since I have not seen such a behavior and we have not heard of other cases so far, looking at the log files seems to be the only way to solve the problem.

While in most cases, problems like this can be related to broken cache files or broken content files, it could of course be hardware related too.

E.g., have you ever tested your RAM with something like https://www.memtest86.com/

I remember a case probably 14 years ago, when a client programmed a show with appr. 10 display computers and all of those did an unintended reboot after some 4-8 hours. Some earlier, some later.

After long investigations, we found out that the RAM modules used in the server didn't work properly when more than one module was used. As the mainboard used quad-channel RAM architecture, there were of course 4 modules in the chassis. Taking three out made the system stable. The RAM was specified in the QVL list of the mainboard!

We then changed the RAM to another product which we used at that time in our servers and that cured the issues.

[Gates kweon]

I finally figured out the cause.

1. It had to do with the resolution.
We used 6 projectors.

However, the recommended resolution of the projector was 1920x1080 for 3 units and
 1920x1200 for 3 units.

Therefore, the output resolution was unified to 1920x1080.

Anyway, the first 4 days were fine.

Production started to shutdown on the 5th day. every 25 minutes

After the event, I tested it on an HD monitor in the office.
There were no problems at all.

The same problem occurred in Useed as well.
Useed is a Korea dealer.
We shared information.

They tested it with different graphics cards and the 3070 didn't have that problem.

No problem with any recommended resolution.

2. This also seems to be a problem with the w.o version. According to them, 
this symptom appeared only in Ver 6.*.* version.

Even if the recommended resolution and the w.o setting resolution are different,
 I think that shutdown should not occur.
Please refer to it in the next version.

 

[artboy]

For resolution, the setting is through Production, but it is the Display server that generates and exports the actual signal.
It's hard to understand why Watchmaker is shut down because of that. 

I've used a mix of WUXGA and FHD in the field, and I've never experienced this symptom. 

I'm testing it in a lab, but it hasn't reproduced the symptoms yet. Can you provide more information?

Is it only when you connect six channels? Or is it the same with fewer channels on the same GPU?

[jfk]

This is baffling. It makes no sense for the resolution settings in the watchmaker stage display specifications to cause watchmaker to crash (shut down). Highly unlikely. Incorrect settings there could cause issues with watchpoint, but not watchmaker. In fact, I am always surprised when erroneous display specifications settings have no negative impact on watchmaker. For example a typo entering a setting, like a resolution of 1921x1080 had no negative effect on watchmaker, but of course watchpoint errored on it while trying to go online.