NAB occasionally freezes

[rdeardorff@spi-ind.com]

NAB occasionally freezes when searching, updating data field mappings, and sometimes just randomly if you are clicking between things too fast

image.png?image.png?
Can you give me pointers on how to find what this is related to?

[rdeardorff@spi-ind.com]

One developer identified this

image.pngnotice the time on getDataSourceWidgets...
?

[robert.swanson]

How many CGI jobs do you have servicing the instance in question?

https://cnxcorp.com/tip-month/bumping-up-the-number-of-cgi-jobs-servicing-your-valence-apps/

[rdeardorff@spi-ind.com]

currently it looks like 2

[rdeardorff@spi-ind.com]

image.png
here is a list of the jobs serviced by this instance?

[robert.swanson]

Looks like your maximum is actually 5 -- the three PGM-jvmStartPa ones, and the two marked PGM-QZSRCGI. That's a pretty low number for anything more than a few simultaneous users.

I would suggest you try bumping it up to 10 or 20, as described in that blog post link.

[rdeardorff@spi-ind.com]

actually we are already set to 40

image.png?

[robert.swanson]

Interesting... Have you ever seen more the 5 cgi jobs spawned?

You might want to double check what you have set in Dynamic Content and CGI > General Settings > Maximum active CGI jobs, ensure it's not restricting the total.

The getDataSourceWidgets call is generally a bit slower because it's doing a lot of heavy lifting, but I wouldn't expect it to take 10 seconds under normal circumstances. More like 2.

[rdeardorff@spi-ind.com]

I have not paid attention to the number of CGI jobs spawned into this instance. as of right now there are 5 but all in TIMW. Looked at other instances and I only see 5 for those as well. On our production system there are 34 spawned. We don't use NAB on production, so we have not seen that error crop up.

Are the number of threads per Valence instance, or are they cumulative across all the instances?

Here are my settings:


image.png?

image.png?

[robert.swanson]

Okay, perhaps the spawned CGI jobs are a red herring then... Let's assume that's all working fine.

Looking back at your original screenshot with the Google Chrome message -- which I think I missed because I was focused on your second post with the long-running calls in the call log -- that's definitely not normal. Is there one particular user for whom this phenomenon with the Chrome pop-up message is happening most often? We'd like that user to try working with Valence in an "Incognito" window for a few days and see if the problem ever happens while in that mode. If the problem does occur again while in Incognito mode, then we'll probably need to get on a web meeting to see what's going on. If it doesn't, then it could indicate some sort of browser plug-in, or something else ancillary to Valence, at play when working in "Normal" mode.

[rdeardorff@spi-ind.com]

Incognito didn't make any difference. Is it possible that AI is taking all the time? anyway to experiment with disabling it? (6.3.20260622)

[robert.swanson]

There's no AI involved in any of those processes.

Appears we need to set up a web meeting with you to see this problem in action on your system. Are you able to reliably replicate this freeze phenomenon?

[rdeardorff@spi-ind.com]

unfortunately it is a random issue with no way to force the error to occur. I have asked the developers to tell me exactly what they are doing (what datasource/widget/app they are edting)....maybe that will help me drill down to the exact issue.