NT8 Crash issue

[crokusek]

Hi!,

Can I please get a programmer to review this stack trace?

https://postimg.org/image/6ut4mggzx/

Request simple interim fix, please just try/catch that section and log it with extended info for next minor release!!! IMO Win/win, done!

Clearly this is not directly related to custom code. It is happening once every few days at random operations and crashes NT8.

Originally reported here: https://ninjatrader.com/support/foru...ad.php?t=90093

I cannot post to that archived thread and Chelsea is not responsive. This is not a dead issue.

Thanks,
Chris

[NinjaTrader_ChelseaB]

Hello Chris,

Currently, I am unable to reproduce this error on my end.

So that I may report this to our development, can you let me know what steps I need to take to reproduce this error?

Are you able to reproduce this error without any custom scripts in the NinjaTrader filesystem?

[crokusek]

This is not easily reproducible (1 every few days) yet it is serious (crash). It might be limited to the unique timing of my hardware because I have only seen it on 1 of 2 machines. I doubt you have 8 hours to sit and play with a workspace either. This makes providing a case moot. I also don't have 8 hours to figure out if it is custom scripts or not. Now if the exception could be caught and improved to at least provide the chart tab name/indicator context then I'd have a place to start creating a test case of the 20 tabs I have open.

Here is where you get hung up for not providing a case...please remain positive toward a solution...

At our company, we capture client exceptions/stack traces and relay them to a sever side database automatically. We then analyze them offline and are able to fix the simpler ones without any need for a test case. This flow does work.

For this case the error message needs improvement anyway, so that means catching it. There should be no holdup here since its a crashing bug. ~10 minutes from a programmer to implement or reject these claims.

Another option, I could debug locally if possible to provide debug/pdb files for that GUI assembly.

Another option, I could call in the next time it happens and have someone remote debug.

[NinjaTrader_Jesse]

Hello,

Thank you for the reply.

You had noted that no custom code was involved to reach this crash, are you currently running a completely stock installation with no imported or custom developed items installed to confirm this result? I noted in the other thread you had used a Custom indicator and were reloading the indicator to create the exception. Can you confirm this happens with any stock indicator or does this only happen with that custom indicator?

If this can occur with any stock item, I would like to collect further details from the installation and the steps used in that case. If you are running a completely clean install, we can exclude the thought of custom items and focus on the platform and computer instead at that point. If there are custom items imported/used during this exception, I would like you to test without those items to check if this is related to custom code.

Unfortunately if the crash only appears infrequently, you may need to further review what you are using in the active workspaces or what actions were completed prior to the crash to find more details. This would be the first step for any remote assistance call on the issue or to see if by reducing what is currently used if the problem still exists. Our support can not provide debugging files nor can we tell development to append items to source based on a stack trace unfortunately. Instead we can assist in working with you to find ways to reproduce what you are seeing to compose a report on what is occurring.


I look forward to being of further assistance.

[crokusek]

Hypothetically say that after a bunch of effort and some luck this case is made reproducible and we transfer the test case to customer support but it turns out that the case will only fail on my machine (because it is a timing issue specific to hardware, say its a "Cray").

From what I understand the issue will be dead to you because you don't have a repro. Can you verify?

If yes, that is just a terrible policy especially for a crash and when there were other means available. In effect, a customer support policy is deciding on behalf of the whole company to be out of the "Cray" market.