Monitoring an ANSYS Fluent Batch Job Remotely
This document describes the methodology to connect and monitor a Rescale ANSYS Fluent batch job remotely using ANSYS Fluent Remote Visualization Client from a Rescale or local Workstation.
- Setup SSH connectivity in Job Settings before any jobs are submitted, as described here.
- A Rescale Batch and Workstation job are required to be submitted
- The Rescale Workstation job can be substituted with a local workstation
- Both the batch job (serverA server is a computer system or software program that provi...) and workstation (client) version of Fluent must be exactly the same
- In this document, we will use ANSYS Fluent 2022 R2.
- The following methodology uses Linux. Windows implementation is also applicable with some modifications.
- This is not supported in the Workbench environment; Fluent stand-alone only.
Submitting a Fluent Batch Job
The general workflow of this step can be followed here ANSYS Fluent Batch Tutorials
In order to start the Fluent Remote Visualization Server when the batch job is submitted, the following syntax is needed to be added in the command box:
An example submit command would look like this:
fluent 3ddp -gu -ssh -cnf=$FLUENT_HOSTS -t$RESCALE_CORES_PER_SLOT -i run_plot.jou -sifile=rescale_remote.txt
If specifying the port range for the server is required, an additional syntax can be added:
A minimum of 2 ports
(portspan=>1) is necessary for a successful connection.
Once the batch job is running, the rescale_remote.txt file will be created in the working directory. It contains the server’s private IP, port number, and Fluent job password. Take note of these as they will be used in the next step.
Connecting from a Rescale Workstation
Create a Workstation job with the ANSYS Interactive Workflow tile. Ensure the same version as the batch job is used.
Connect to the GUI interface, click Manage Desktop on the desktop, and select the batch job submitted in the previous step.
rescale_remote.txt in the Live Tailing window, and download the file.
Setup a reverse SSH tunnel to the batch job from a terminal window (assuming ssh keys are implemented correctly).
rescale_remote.txt file contains an IP address and port (shown above), these should be included in the following command after the
-L flag, e.g.:
ssh -i id_rsa -p 22 udeprod_SBLKd@18.104.22.168 -L 46249:10.29.3.246:46249 -v -N
In this case:
|id_rsa||Private key file for SSH|
|22 (or 32100)||SSH connection port|
|udeprod_SBLKd||Batch job username|
|22.214.171.124||Batch job public IP|
|46249||Fluent Remote Viz server port*|
|10.29.3.246||Fluent Remote Viz server private IP|
Find rescale_remote.txt downloaded previously and edit it, exchanging the IP with:
The final file should look something like this:
Running ANSYS Fluent Visualization Client
From a new terminal window, run the command:
When the client is loaded, select the modified
Click the Connect button and the client will successfully connect to the server.
Check the Host’s IP address matches the server’s private IP address and the Case is the job running on the remote batch job.
Connecting from a Local Workstation
The procedure is the same as connecting from a Rescale Workstation. Ensure the required ports are open on the firewall and if necessary use the following in the batch job command to specify the ports that are allowed.
It may also be necessary to modify the SSH tunneled port to an open port. To check the necessary port is open, use the command:
ss -tupln |grep [port number]
This should return
[port number] is
To check for connectivity, use the following command:
nc localhost [port number]
This should return a non-negative response.
Using ANSYS Fluent Remote Visualization Client
Up to six Remote Visualization Sessions can be made in one instance of the application. No ANSYS license is required for this application.
It is possible to interact with the batch job calculation, by pausing and making changes.
Caution: Pressing Interrupt will STOP the Rescale batch job. There is NO “Are You Sure?” prompt to confirm your action.
It is also possible to show intermediate plots that have been predefined in the batch job.