Bokeh crashing Jupyter over the past two days/running super slow

Hey,

any idea why Bokeh may be either crashing my Jupyter notebooks or running snail slow over the past two days? I haven’t had any problems before. I’m using fairly big DataFrames so I’m guessing that may have something to do with it?

Both Bokeh and the Notebook are large and complicated and integrating them together is often a challenge. It's hard to speculate without more information, and without a way to reproduce the problem. Can you share a notebook and some data that reproduce what you are seeing?

Thanks,

Bryan

···

On Dec 5, 2016, at 11:07 AM, Kasia Rachuta <[email protected]> wrote:

Hey,

any idea why Bokeh may be either crashing my Jupyter notebooks or running snail slow over the past two days? I haven't had any problems before. I'm using fairly big DataFrames so I'm guessing that may have something to do with it?

--
You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/171baa93-7f6f-4cd4-ab52-f173e3703a15%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Hey Bryan,

For confidentiality purposes, I unfortunately cannot post anything :frowning: The one thing I know for sure is that the exact same code worked last week - I’ve now re-run it this week and it’s crashing… Yesterday afternoon some of it worked. I suppose I will just keep on trying.

Kasia

···

On Monday, 5 December 2016 09:11:42 UTC-8, Bryan Van de ven wrote:

Both Bokeh and the Notebook are large and complicated and integrating them together is often a challenge. It’s hard to speculate without more information, and without a way to reproduce the problem. Can you share a notebook and some data that reproduce what you are seeing?

Thanks,

Bryan

On Dec 5, 2016, at 11:07 AM, Kasia Rachuta [email protected] wrote:

Hey,

any idea why Bokeh may be either crashing my Jupyter notebooks or running snail slow over the past two days? I haven’t had any problems before. I’m using fairly big DataFrames so I’m guessing that may have something to do with it?


You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/171baa93-7f6f-4cd4-ab52-f173e3703a15%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Is there something you can identify as having changed? library or notebook or browser versions?

Bryan

···

On Dec 6, 2016, at 10:10 AM, Kasia Rachuta <[email protected]> wrote:

Hey Bryan,

For confidentiality purposes, I unfortunately cannot post anything :frowning: The one thing I know for sure is that the exact same code worked last week - I've now re-run it this week and it's crashing... Yesterday afternoon some of it worked. I suppose I will just keep on trying.

Kasia

On Monday, 5 December 2016 09:11:42 UTC-8, Bryan Van de ven wrote:
Both Bokeh and the Notebook are large and complicated and integrating them together is often a challenge. It's hard to speculate without more information, and without a way to reproduce the problem. Can you share a notebook and some data that reproduce what you are seeing?

Thanks,

Bryan

> On Dec 5, 2016, at 11:07 AM, Kasia Rachuta <[email protected]> wrote:
>
> Hey,
>
> any idea why Bokeh may be either crashing my Jupyter notebooks or running snail slow over the past two days? I haven't had any problems before. I'm using fairly big DataFrames so I'm guessing that may have something to do with it?
>
> --
> You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
> To post to this group, send email to [email protected].
> To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/171baa93-7f6f-4cd4-ab52-f173e3703a15%40continuum.io.
> For more options, visit https://groups.google.com/a/continuum.io/d/optout.

--
You received this message because you are subscribed to the Google Groups "Bokeh Discussion - Public" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To post to this group, send email to [email protected].
To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/5afccc51-66b5-441c-85f4-4f4c01498f88%40continuum.io.
For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Literally nothing I can think of…

···

On Tue, Dec 6, 2016 at 8:15 AM, Bryan Van de Ven [email protected] wrote:

Is there something you can identify as having changed? library or notebook or browser versions?

Bryan

On Dec 6, 2016, at 10:10 AM, Kasia Rachuta [email protected] wrote:

Hey Bryan,

For confidentiality purposes, I unfortunately cannot post anything :frowning: The one thing I know for sure is that the exact same code worked last week - I’ve now re-run it this week and it’s crashing… Yesterday afternoon some of it worked. I suppose I will just keep on trying.

Kasia

On Monday, 5 December 2016 09:11:42 UTC-8, Bryan Van de ven wrote:

Both Bokeh and the Notebook are large and complicated and integrating them together is often a challenge. It’s hard to speculate without more information, and without a way to reproduce the problem. Can you share a notebook and some data that reproduce what you are seeing?

Thanks,

Bryan

On Dec 5, 2016, at 11:07 AM, Kasia Rachuta [email protected] wrote:

Hey,

any idea why Bokeh may be either crashing my Jupyter notebooks or running snail slow over the past two days? I haven’t had any problems before. I’m using fairly big DataFrames so I’m guessing that may have something to do with it?

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/171baa93-7f6f-4cd4-ab52-f173e3703a15%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/5afccc51-66b5-441c-85f4-4f4c01498f88%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/1758E5A0-611E-440A-BC28-5B1653EB3468%40continuum.io.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

You can try clearing the outputs of the notebook itself, by loading the notebook, clearing out all the outputs, saving it, then closing the tab, and then restarting the jupyter server, and opening the notebook again. This will remove all embedded javascript and HTML elements in output cells of the notebook, which might be causing issues.

You can also see if the code is saving out to files or somehow “accumulating” data or state over the course of runs…

If it’s not doing that, but it’s making a query against a “live” database or some such, then it’s possible that the query result set has unexpectedly grown large over the last few days? Have you tried doing some data sanity checks to ensure that dataset sizes are what you expect?

HTH,

Peter

Hi Peter,

Thanks for the advice. I actually copied and pasted the code into a new Jupyter notebook, re-run it and the same thing happened! I just checked and none of the tables have gotten that much bigger (yes, I’m running a SQL query first). I’m really stuck and gutted - I have a meeting tomorrow and was planning on showing all the visuals then…

Kasia

···

On Tue, Dec 6, 2016 at 8:38 AM, Peter Wang [email protected] wrote:

You can try clearing the outputs of the notebook itself, by loading the notebook, clearing out all the outputs, saving it, then closing the tab, and then restarting the jupyter server, and opening the notebook again. This will remove all embedded javascript and HTML elements in output cells of the notebook, which might be causing issues.

You can also see if the code is saving out to files or somehow “accumulating” data or state over the course of runs…

If it’s not doing that, but it’s making a query against a “live” database or some such, then it’s possible that the query result set has unexpectedly grown large over the last few days? Have you tried doing some data sanity checks to ensure that dataset sizes are what you expect?

HTH,

Peter

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/CABGrp%2B9TcPe315qwrb_2gtjEYF-JtDstKN%3DU8L_VJFr24CpKGQ%40mail.gmail.com.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

I forgot to add - not sure if that helps. I’m producing a bunch of histograms over a grid plot. I can run the code that outputs the plot (doesn’t show it) and it’s only when I run “p = gridplot(hist_period1, hist_period2)
show§”, that it crashes.

···

On Tue, Dec 6, 2016 at 8:38 AM, Peter Wang [email protected] wrote:

You can try clearing the outputs of the notebook itself, by loading the notebook, clearing out all the outputs, saving it, then closing the tab, and then restarting the jupyter server, and opening the notebook again. This will remove all embedded javascript and HTML elements in output cells of the notebook, which might be causing issues.

You can also see if the code is saving out to files or somehow “accumulating” data or state over the course of runs…

If it’s not doing that, but it’s making a query against a “live” database or some such, then it’s possible that the query result set has unexpectedly grown large over the last few days? Have you tried doing some data sanity checks to ensure that dataset sizes are what you expect?

HTH,

Peter

You received this message because you are subscribed to the Google Groups “Bokeh Discussion - Public” group.

To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].

To post to this group, send email to [email protected].

To view this discussion on the web visit https://groups.google.com/a/continuum.io/d/msgid/bokeh/CABGrp%2B9TcPe315qwrb_2gtjEYF-JtDstKN%3DU8L_VJFr24CpKGQ%40mail.gmail.com.

For more options, visit https://groups.google.com/a/continuum.io/d/optout.

Kasia,

    I would recommend

trying to create a virgin Conda environment and run your
Notebook from there. Something like the following might be a way
to achieve this

conda create -n kasia anaconda=4.2 python=MAJOR.MINOR # replace with your preferred version: 2.7 or 3.5 presumably
source activate kasia
jupyter notebook

    Then inside your

notebook make sure you’re running a Python kernel that is from
the kasia
Conda environment by executing:

import sys
sys.prefix

    And finally, before

you run the troublesome gridplot()
function, make sure you don’t have any memory hogs on the
system. On Mac you should use Activity Monitor, on Windows you
can use Task Manager and on a Unix-variant you can use “top” and
then sort by memory consumption.

    If that is

inconclusive you might want to reboot the system if you’re able
to.

    Then run your code,

including the gridplot()
function.

Regards,

Ian