Ca2+ dynamics (Cluster) MATLAB Error

[AnnaSophia22]

Dear developer,
I am working on a setup to model the Calcium dynamics of Astrocytes with Astro.
I am following the 'Ca2+ dynamics (Cluster)' instructions in 'Astro User guide v7'. I am trying to run the setup on a single host (Windows 11, MATLAB 2023a, NEURON, Astro 1.0. Furthermore, I have set up a worker in a Linux guest (Ubuntu 20.04, NEURON installed via pip3) running in Virtualbox with host/guest networking enabled (both devices can ping each other and win can ssh into guest).

When I Astro's './Start.m' (pressing 'Ca2+ dynamics' button) or './cluster/host/Start_Simulation.m' and select item '1. Start simulation from scratch.', then I get the following error:

SWITCH expression must be a scalar or a character vector.

Error in UpdateViewControls>UpdateYForNameAndUnitLabels (line 156)
switch style

Error in UpdateViewControls (line 90)
UpdateYForNameAndUnitLabels(handler, style, j, panIdx_);

Error in UpdateView (line 4)
UpdateViewControls();

Error in PrepareAndShowGUI>figure_ResizeFcn (line 191)
UpdateView();

Error in PrepareAndShowGUI (line 104)
figure_ResizeFcn();

Error in START_Simulation (line 64)
PrepareAndShowGUI();

Any feedback or suggestions would be greatly appreciated.

Best regards,
AnnaSophia

[AnnaSophia22]

I have to add information: When I Astro's './Start.m' (pressing 'Ca2+ dynamics' button), I get exactly the following output:

Checking whether HPC kernel is running in background mode ...
Checking annasophia ...
Checking as1 ...
It is not.
Checking whether results of simulation are present in HPC kernel directory ...
Please specify what to do:

1. Start simulation from scratch.
2. Do nothing
   (If i select 1 I get - obviously - similar Errors mentioned above.)

Can anybody please help me to find the root of this problem?
Is there maybe a problem with my headnode (annasophia) or my slave (as1). I specified them uniformly in the 'params.bat', BasicParams.m, hostfile_BusyMaster, hostfile_IdleMaster, therefore I don't know how I could fix the issue...

Hoping for a helpful reply,
Anna Sophia

[LeonidSavtchenko]

Hello, From the information you provided, it seems that there might be an issue with the HPC kernel or the specified hosts in the 'params.bat', BasicParams.m, hostfile_BusyMaster, hostfile_IdleMaster. In order to get a better understanding of the issue, could you please provide me with more information such as the version of Astro you are using, the operating system, and any other relevant details? Once I have received this information, I can begin to investigate the issue further and provide you with a solution. Now here the docker file ******************************** FROM ubuntu:16.04 RUN apt-get update && apt-get install -y openssh-server libopenmpi-dev g++ libxt-dev libncurses-dev make unzip RUN mkdir /var/run/sshd RUN echo 'root:screencast' | chpasswd #install neuron RUN wget -P /tmp https://neuron.yale.edu/ftp/neuron/versions/v7.6/7.6.2/nrn-7.6.2.tar.gz RUN tar -C /tmp/ -xvf /tmp/nrn-7.6.2.tar.gz RUN mkdir /usr/local/neuron RUN mv /tmp/nrn-7.6 /usr/local/neuron/nrn WORKDIR /usr/local/neuron/nrn RUN ./configure --prefix=`pwd` --with-paranrn --without-iv RUN make install # Download astro from GitHub WORKDIR /ucl RUN pwd \ && wget https://github.com/LeonidSavtchenko/Astro/archive/master.zip \ && unzip master.zip # SSH login fix. Otherwise user is kicked off after login RUN sed ***@***.******@***.*** optional ***@***.***' -i /etc/pam.d/sshd ENV NOTVISIBLE "in users profile" RUN echo "export VISIBLE=now" >> /etc/profile COPY astro.pub /root/.ssh/ RUN useradd ucl -m -s /bin/bash \ && echo 'ucl:ucl' | chpasswd \ && usermod -aG sudo ucl \ && echo "AllowUsers ucl" >> /etc/ssh/sshd_config \ && chown -R ucl:ucl /ucl EXPOSE 22 COPY docker-entrypoint.sh / ENTRYPOINT ["/docker-entrypoint.sh"] CMD ["/usr/sbin/sshd", "-D"] ******************************** Best regards, Leonid

…

On Fri, 26 Jan 2024 at 23:03, AnnaSophia22 ***@***.***> wrote: I have to add information: When I Astro's './Start.m' (pressing 'Ca2+ dynamics' button), I get exactly the following output: Checking whether HPC kernel is running in background mode ... Checking annasophia ... Checking as1 ... It is not. Checking whether results of simulation are present in HPC kernel directory ... Please specify what to do: 1. Start simulation from scratch. 2. Do nothing (If i select 1 I get - obviously - similar Errors mentioned above.) Can anybody please help me to find the root of this problem? Is there maybe a problem with my headnode (annasophia) or my slave (as1). I specified them uniformly in the 'params.bat', BasicParams.m, hostfile_BusyMaster, hostfile_IdleMaster, therefore I don't know how I could fix the issue... Hoping for a helpful reply, Anna Sophia — Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYAKA2KBFCL6F6TE6E4BNLYQQY2JAVCNFSM6AAAAABBBPGIUGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMJSHAYDOMRVGQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[AnnaSophia22]

Dear Mr. Savtchenko,

I'd be glad to tell you my set up:

Worker: Ubuntu 20.04, mpiexec (OpenRTE) 4.1.2, Neuron 8.2.3,
Worker/Host connection: VirtualBox 7.0 with a shared Folder
Host: Windows 11, Matlab R2023a, Neuron 8.2.0+, MPI 10.1.12498.18

Relating to the docker file I have some problems with astro.pub and docker-entrypoint.sh. Shold they exist in the build context?

Moreover I ran the dockerfile without the last lines (i deleted all lines below COPY astro.pub ...) and I got the following error:

I would be very grateful for any help.

Best regards,
Anna Sophia

[LeonidSavtchenko]

Dear Anna Sophia, Thank you for providing your setup information. It seems that you are experiencing some issues with the docker file and the files astro.pub and docker-entrypoint.sh. These files should indeed exist in the build context. Regarding the error you encountered when running the docker file without the last lines, could you please provide more details or the specific error message? This will help me better understand the problem and provide you with a more accurate solution. In the meantime, I have attached a Word file that describes how to use Docker containers and Docker for Amazon Web Services (AWS). I hope this information will be helpful to you. Please let me know if there is anything else I can assist you with. cheers L

…

On Wed, 7 Feb 2024 at 00:36, AnnaSophia22 ***@***.***> wrote: Dear Mr. Savtchenko, I'd be glad to tell you my set up: Worker: Ubuntu 20.04, mpiexec (OpenRTE) 4.1.2, Neuron 8.2.3, Worker/Host connection: VirtualBox 7.0 with a shared Folder Host: Windows 11, Matlab R2023a, Neuron 8.2.0+, MPI 10.1.12498.18 Relating to the docker file I have some problems with astro.pub and docker-entrypoint.sh. Shold they exist in the build context? Moreover I ran the dockerfile without the last lines (i deleted all lines below COPY astro.pub ...) and I got the following error: grafik.png (view on web) <https://github.com/LeonidSavtchenko/Astro/assets/154735915/ec7d7b98-5081-4d36-adc7-10f2cb9bee79> I would be very grateful for any help. Best regards, Anna Sophia — Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYAKAZJHVOPL35HYRIQNL3YSLEAXAVCNFSM6AAAAABBBPGIUGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZRGAZDKNJVGQ> . You are receiving this because you commented.Message ID: ***@***.***>

[AnnaSophia22]

Dear Mr. Savtchenko,

I solved the issues through installing putty (https://www.chiark.greenend.org.uk/~sgtatham/putty/latest.html), because it is necessary that plink and pscp run in the background. Moreover I installed MatlabR2017a, because Astro1.0 had troubles with MatlabR2023a.
I had also a problem with the path of nrniv in run.sh in ClusterCaSim-hpc-scripts. In worker/linux the path had been correct whearby in host/windows there where wrong paths, but maybe this is just a problem of my individual setup.

Now the Ca-simulations run and I no longer receive error outputs from the Matlab file. Nevertheless I don't get the wanted simulation plots: If I run the Start.m Matlab file and press the "NEURON Simulations" button, the wanted new windows appear. If I press the "Select diameter distribution for nano geometry" button, load the "testshape.dat_radii_dist.txt" file and click on the "Start Astro." button the window closes (the other two windows stay open) and the process terminates.

Due to the user guide there should be opened through this way the main control panel (shown below), but I don't get there. Therefore I'd like to ask: Where/how can I search for issues? If I should search in the .hoc files of the "NeuronSims" folder, how can I set breakpoints there/run the files step by step and in which file or where exactly is the main control panel created?

I'd appreciate any help and send kind regards,
Anna Sophia

[AnnaSophia22]

Dear Mr. Savtchenko,

I am unfortunately still stuck at the same error. May I please ask you with which exact version of NEURON the full version of Astro 1.0 is running on your computer?

Because the Astro-Paths seem often not to match with the Paths of the nrn environment due to differences of the folder orders between the various NEURON versions. (I tested the current version NEURON 8.2 but the paths do not match at all, NEURON 7.2 but it contains no mingw folders, NEURON 7.3 but it did not contain the bin folder, now I run NEURON 7.4 but it does not find "x86_64-w64-mingw32-gcclib"

I look forward to receiving your reply.

Yours sincerely,
Anna Sophia Preindl

[LeonidSavtchenko]

Hi, Anna I'll look into it again and try to figure it out. However, I want to let you know that we have been using a newer version called BrainCell, which you can download from https://github.com/RusakovLab/BrainCell. The best thing about this version is that it allows you to save the compute module and use the neuroscience cluster for free from https://www.nsgportal.org/overview.html. BrainCell is a much better version, and it can handle any task you need to compute. As I am currently working with a new version, I can assist you better. Maybe you could give it a try? The Astro function of connecting with the cluster is good, but it can be challenging to implement. Nevertheless, I will still check the Neuron version. cheers L

…

On Mon, 22 Apr 2024 at 20:02, AnnaSophia22 ***@***.***> wrote: Dear Mr. Savtchenko, I am unfortunately still stuck at the same error. May I please ask you with which *exact version of NEURON* the full version of Astro 1.0 is running on your computer? Because the Astro-Paths seem often not to match with the Paths of the nrn environment due to differences of the folder orders between the various NEURON versions. (I tested the current version NEURON 8.2 but the paths do not match at all, NEURON 7.2 but it contains no mingw folders, NEURON 7.3 but it did not contain the bin folder, now I run NEURON 7.4 but it does not find "x86_64-w64-mingw32-gcclib" I look forward to receiving your reply. Yours sincerely, Anna Sophia Preindl — Reply to this email directly, view it on GitHub <#4 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AEYAKAYLWXCVTU2A2JLU4FLY6VGC7AVCNFSM6AAAAABBBPGIUGVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDANZQGQ3DCNJTHA> . You are receiving this because you commented.Message ID: ***@***.***>