bokeh: Multiple output_file in a single process can fail when models are shared

I have 2 anaconda environments in Windows 7:

• env1 has jupyter lab 0.32 and bokeh 0.12.16,
• env2 has jupyter lab 2.0.1 and bokeh 2.0.0.

In a jupyter notebook, I have 2 cells of code:

Cell 1

source1 = ColumnDataSource(fertility_df[fertility_df['Continent']=='ASI'])
p1 = figure(x_axis_label='fertility (children per woman)', y_axis_label='female literacy (% population)')
p1.circle('fertility', 'female literacy', source=source1)

source2 = ColumnDataSource(fertility_df[fertility_df['Continent']=='EUR'])
p2 = figure(x_axis_label='fertility (children per woman)', y_axis_label='female literacy (% population)')
p2.circle('fertility', 'female literacy', source=source2)

source3 = ColumnDataSource(fertility_df[fertility_df['Continent']=='AF'])
p3 = figure(x_axis_label='fertility (children per woman)', y_axis_label='female literacy (% population)')
p3.circle('fertility', 'female literacy', source=source3)

source4 = ColumnDataSource(fertility_df[fertility_df['Continent']=='LAT'])
p4 = figure(x_axis_label='fertility (children per woman)', y_axis_label='female literacy (% population)')
p4.circle('fertility', 'female literacy', source=source4)

row1 = [p1, p2]
row2 = [p3, p4]
layout = gridplot([row1, row2])
output_file('bokeh.html')
show(layout)

Cell 2

tab1 = Panel(child=p1, title='Latin America')
tab2 = Panel(child=p2, title='Africa')
tab3 = Panel(child=p3, title='Asia')
tab4 = Panel(child=p4, title='Europe')

layout = Tabs(tabs=[tab1, tab2, tab3, tab4])
output_file('bokeh.html')
show(layout)

In the env1, I could run cell 1, cell 2, and cell 1 again without any problem.

However, in env2, after executing cell 1, then trying to run cell 2, I got an error :

RuntimeError: Models must be owned by only a single document, WheelZoomTool(id='1130', ...) is already in a doc

My questions are:

• Is it a bug in bokeh 2.0.0?
• What does it mean by Models in the error message?