Jupyter freezes, needs forced quit, in Prof. Ed. 2019.1.3.

While working in the editor side of Jupyter (1.0.0 via Anaconda3), the IDE keeps freezing (with Apple's "spinning beachball of death"). 

OSX 10.14.5 / PyCharm 2019.1.3 / Python 3.7 / Anaconda3 / Jupyter 1.0.0 (core 4.5)

This is happening very frequently, on average I'd say about once per hour.

It's not the only problem I have been having with the new Jupyter/PyCharm. JetBrains people what is going on ??

This is very detrimental to my productivity (which is the opposite reason I chose PyCharm and the professional edition).

 

17 comments
Comment actions Permalink

Yes. PyCharm goes up about 780% CPU

Reading about the related issue, I am also using matplotlib, but I need to use this library. 

Also I have jdk1.8.0_201.jdk installed. (I am using PySpark which apparently prefers version 8).

  • I have the logs now but I can only see how to attach an image or insert a link here. How should I attach the log files?
  • I was not able to locate the CPU snapshot. I tried to follow this: "Invoke Help | Diagnostic | Profile Indexing. Click Invalidate and Restart (please note that Local History and Indices will be destroyed). Click Open Snapshot in ... to locate the snapshot." I selected "Profile indexing" from PyCharm's Help | Diagnostic dropdown menu. I selected "Invalidate caches / Restart" from PyCharm's File dropdown menu. It did stuff, CPU use went up sharply. But I don't know what/where is "Open Snapshot in ..." 
0
Comment actions Permalink

The key part here is JDK 8

Unfortunately, Jupyter support in PyCharm 2019.x requires JDK 11, no way around it.

Also, I don't see why you can't run PyCharm under JDK 11 and PySpark under JDK 8. From what I've just googled, PySpark use Java from JAVA_HOME env variable. But PyCharm has other ways to select boot JDK, described here: https://intellij-support.jetbrains.com/hc/en-us/articles/206544879-Selecting-the-JDK-version-the-IDE-will-run-under

So you can run PyCharm with JDK 11, while JAVA_HOME will be pointing to JDK 8, which will be picked up by PySpark. There may be other ways to configure JDK for PySpark too.

0
Comment actions Permalink

First, I set the JAVA_HOME variable to 1.8 as your comment suggests.

Next, I looked into switching this "Boot JDK" you mentioned. However, as you can see from the screenshot here, it seems it is already configured to version 11.This JDK is located in /Applications/PyCharm.app/Contents/jdk/Contents/Home/bin and when I do ./java -version, I get openjdk version "11.0.2" 2019-01-15. (As opposed to when I do java -version, which gives me that brand new Oracle jdk that I don't now think I needed to install: java version "11.0.3" 2019-04-16 LTS.

So it looks like the required version of JDK (version 11) comes bundled with and pre-configured as the Boot JDK with version PyCharm 2019. 

The key part here is that the reason for Jupyter-PyCharm 2019.1.3 freezing/crashing and burning my CPU up, remains unsolved.

0
Comment actions Permalink

Yes, JDK 11 is bundled with PyCharm, but you can select to use any other JDK via various methods. Anyway, check Help > About - what JDK version it shows? This is JDK actually used by PyCharm. If it's JDK 11 then the issue is something else.

In that case, please try to open and edit any Markdown (.md) file - do you get performance issue? (Markdown and Jupyter are using the same components).

0
Comment actions Permalink

Yes the boot JDK is 11, as I've made explicitly clear in my previous post with massive screenshot.

You still have not told me how to send you the logs or how to find the cpu snapshot.

 

I'm aborting use of PyCharm-Jupyter as the product was clearly not ready for release and JetBrains support is ineffective.

 

0
Comment actions Permalink

I'm mostly interested what do you have in Help > About pop up, as this reliably shows which JDK is currently used by PyCharm. Then, after confirming your current boot JDK, I want to you to try creating and editing any Markdown file, and if the issue is also reproduced, it will be much easier to identify.

You can upload the logs and CPU snapshots to: https://uploads.services.jetbrains.com/ and provide file name here.

0
Comment actions Permalink

Hi, I am having the exact same problem. Attached is my screenshot of About:

Will follow up with logs as asked to previous submitter.

Also I can't access the issue you gave in your first reply.

This is a very annoying bug as I also crash about once an hour, regardless of which Mac PyCharm is installed on. This has happened for at least a year.

1
Comment actions Permalink

Hi, 

 

The ticket has restricted access, as there is sensitive information attached.  Please feel free to create a ticket here https://youtrack.jetbrains.com/issues/PY  with logs attached. 

0
Comment actions Permalink

This is still happening.  Any use of the Jupyter plugin and PyCharm has to be force-quit several times per hour after it suddenly spikes system CPU.  This renders the product nearly unusable and highly disruptive.  JetBrains support thought it may be caused by the bundled Markdown Support plugin (which you can't/shouldn't disable) and recommended renaming the PyCharm folder at /Users/<user>/Library/Application Support/PyCharm2019.3/  --> "PyCharm2019.3_ old".  This then requires that you reinstall all your plugins, which is a pain, but at least it stopped the problem temporarily.  However, after a day or two the problem is back.  Is there any sort of adequate fix planned for this issue?!

0
Comment actions Permalink

Having the same issue..

 

PyCharm 2019.3.4 (Professional Edition)
Build #PY-193.6911.25, built on March 18, 2020
Licensed to XXXXXXXXXXXXX
You have a perpetual fallback license for this version
Subscription is active until December 12, 2020
Runtime version: 11.0.6+8-b520.43 x86_64
VM: OpenJDK 64-Bit Server VM by JetBrains s.r.o
macOS 10.14.6
GC: ParNew, ConcurrentMarkSweep
Memory: 4029M
Cores: 8
Registry:
Non-Bundled Plugins: DeepBugsPython, com.jetbrains.ChooseRuntime, com.leinardi.pycharm.pylint, net.ashald.envfile, net.seesharpsoft.intellij.plugins.csv, org.intellij.RegexpTester, ru.adelf.idea.dotenv, ru.meanmail.plugin.pyannotations, ru.meanmail.plugin.requirements

0
Comment actions Permalink

Antonina Belianskaya I have created the issue @ https://youtrack.jetbrains.com/issue/PY-41792 with my logs.

I have started a CPU usage profiling in the hope it survives the crashes. This is happening every hour now.

0
Comment actions Permalink

I am having the same issue with version 2020.1.1

I literally bought PyCharm Professional Edition for its Jupyter support.   I'm quite disappointed that this issue seems to have persisted without a clear solution, and will switch to VS Code in the meantime.

 

PyCharm 2020.1.1 (Professional Edition)
Build #PY-201.7223.92, built on April 30, 2020
Licensed to ############
Subscription is active until February 24, 2021
Runtime version: 11.0.6+8-b765.40 x86_64
VM: OpenJDK 64-Bit Server VM by JetBrains s.r.o
macOS 10.15.4
GC: ParNew, ConcurrentMarkSweep
Memory: 725M
Cores: 16
Non-Bundled Plugins: R4Intellij, aws.toolkit

0
Comment actions Permalink

@Ltemery

I'm sorry for the inconvenience. 2020.2 is planned to be the version where Jupyter preview will be migrated to JCEF instead of JavaFX. This change expected to fix Jupyter performance issues. 

0
Comment actions Permalink

When is that going to be ready?   This is literally non functional. It freezes every few lines, so I've had to stop using Pycharm.

As I mentioned, the Jupyter functionality is the sole reason that I upgraded to PyCharm Professional.  The morally correct business choice for having a non-functional advertised capability would be to not charge me for any of the time before you introduce the new fix.

I will be moving to VS Code for the time being, 

0
Comment actions Permalink

This is soul-crushing. The ONLY reason I bought professional, while being berated for doing so by my colleagues who are better coders than me, was so I could sit in front of your awesome IDE and debug my Jupyter sessions .

I haven't seen so many beachballs of death since ... no can't even remember another instance with any other piece of software. Worst of all Jupyter integration is one of the most heavily advertised features of 2020 Pro

Please get this fixed asap with 2020.2. If not, there is zero chance I will renew my account and the whole experience will just suck because I had such high hopes. As of now is totally unusable on two machines:

Macbook Pro 2018 Mojave - 16gig

iMac 2019 Catalina - 40 gig

 

0
Comment actions Permalink

You can try 2020.2 EAP right now: https://www.jetbrains.com/pycharm/nextversion/

The issues should be fixed there.

0

Please sign in to leave a comment.