ccu, Administrators
684
edits
Changes
Jump to navigation
Jump to search
m
→Moving your workloads to the new cluster
Note that there are timeouts in place - this is a demo pod which runs only for 24 hours and an interactive session also has a time limit, so it is better to build a custom run script which is executed when the container in the pod starts. See the Kubernetes documentation on pods and jobs for more details. TODO: link to respective doc.
'''Important known issue: Currently, there seem to be problems with running Tensorflow applications on at least one of the compute nodes (node zariel), the nVidia GPU device is not always created as "/dev/nvidia0", but sometimes as "/dev/nvidiaX", where X is another number (likely the device ID on the host). This makes Tensorflow fail to detect Please use tiamat in the GPUmeantime. While I am searching for an actual solution, a workaround is'''<syntaxhighlight>(shell inside the container)# ln -s /dev/nvidiaX /dev/nvidia0</syntaxhighlight>
If you do not have your code ready, you can do a quick test if GPU execution works by running demo code from [https://github.com/dragen1860/TensorFlow-2.x-Tutorials this tutorial] as follows:
