I lose my data every time I stop my pod
Could somebody please provide a simple explanation for how I'm supposed to use network storage for persistent data? Every time I stop a pod I've been using it becomes unavailable next time, so I have to choose a different GPU on a different datacenter. When the pod is launched it does not connect to my network storage on the original datacenter, so I have to start everything all over again. The whole datacenter/pod thing is incredibly confusing
3 Replies
1. you need to create a network storage volume. https://www.runpod.io/console/user/storage
a. pick the location which has the gpus you intend to use. this is important, because your network storage has to be sited in the same location as the gpu you will want to use.
Create a network volume | RunPod Documentation
Create a network volume in Secure Cloud to access high-performance storage and flexibility for multiple pods, with options for data center selection, naming, and size allocation, and enjoy cost-effective storage solutions with robust infrastructure and NVME SSDs.
b. once that's done. your network storage will appear in https://www.runpod.io/console/user/storage. click deploy and and it will bring you to launch a pod that will have access to your network storage
c. once you launch. follow your normal way to access you pod. once in you can find your network storage under the workspace directory (i.e. cd workspace).
NOTE: if you do not deploy a pod that is connected to your network storage, there will still be a workspace directory. but it will not be persistent.
to check, you can look at your pod instance on the UI https://www.runpod.io/console/pods and if it has the field "volume:" and your network storage name, it means its connected.
hope this helps!