Forums

Kernel restart failed and then dead kernel

Does anyone facing "Dead Kernel" problem in python notebook? It happens to me quite often that once it starts "Dead Kernel", it will keep restart and die and restart .... Even if I start a new python notebook, same thing happens.

In the console page, I found there are many running process and if I kill them all, the python notebook will back to normal. However, it is frustrated that i have this problem every hour.

Can anyone help?

That's a little strange, thanks for reporting it. We'll keep an eye out for it when we next do some updates to the notebook service...

DO you have any idea what could be causing the dead kernel? I don't think it's something we've seen ourselves, so maybe something about the kind of work you're doing with the notebook?

I will work out a sample for you. Please wait and keep you post.

I tried and cannot reproduce the problem. It seems like there is no particular orders or files to make it dead. However, right now, my ipython notebook is dead and every ipynb files i start will make the kernel goes dead.

I attached the process list. Is there any information you need to help your investigation?

my process list

It looks like you have many kernels running. If you're sure that they are not being used, you should kill them.

Yes. That's how i solve the problem by killing all the dead process. But i am not sure how I run into this problems. It is quite annoying that it happens all the time.

Then my guess is that one of your notebooks is doing something that causes some issue with the notebook server and that is causing the kernels of the others to be inaccessible. Try to see if that is the case by running your notebooks one-by-one in different orders to see if there is one that causes the issue.

Thanks for your suggestion.

Just curious, when I done with my ipython notebook and i want to close it, i usually click "back" in my chrome browser. After a while, I check the process list and the process is there. As time goes by, more and more processes are created and will not be killed. When there are about 15-20 process, the kernel dead happens.

Is there a proper way to close a notebook?

Looks like we have a bug that we'll need to fix. In the meantime, if you need to close all your kernels, you can go into the process list and find the process that has jupyterhub-singleuser in it's name. If you kill that, it will kill all your kernels.

Thanks for your advice and I will wait for your bug fixing. Really appreciate your quick response.

Hi, Im testing a Notebook , yesterday was working ok but now is crushing (I didnt do any changes) . I already killed the services but still is crushing. What can i do ? Can you help me with that? Can i give you the path of the notebook and can you check it ? Thanks in advance

Hi, can you give us more info on how it is crashing? eg: what happens when you are trying to do x?