Looks like I keep having to revisit how I'm running my WSL2 instances. Previously, I had needed to create an ext4-formatted VHD to host PVCs within (first) KinD and (then) Minikube clusters. At the time that I wrote my (KinD) articles, I only had a single WSL2 instance (running Oracle Enterprise Linux 9). As such, the wsl-level mount command that I was using seemed to be sufficient. However, as my needs evolved, I needed to be able to run more than one WSL2 instance. I noticed that, even though my mount command was succeeding, the mount point was no longer showing up inside any of my instances. Digging around, it seems like, because my VHD-hosted ext4 filesystem was created on the bare disk, WSL stopped being able figure out where I wanted it mounted inside my (preferred) WSL2 instance. Turns out, the fix was to give the mount command a bit more information. Instead of the (now insufficient):
wsl --mount --name openclaw-data --vhd G:\OpenClaw\openclaw-data.vhdx
I need to update the command to
wsl --mount --name openclaw-data --vhd G:\OpenClaw\openclaw-data.vhdx --type ext4
Adding that `--type ext4` seems to have (for now, at least) given WSL2 enough additional information about my intentions that, when I log into my (preferred) WSL2 instance, it now shows up where expected:
$ df -Pht ext4 Filesystem Size Used Avail Use% Mounted on /dev/sdd 251G 32G 207G 14% / /dev/sde 118G 65G 49G 58% /mnt/wsl/openclaw-data
Honestly, I don't know why the first one ever worked: "dumb-luck", presumably. Hopefully, I won't have to continue tweaking these kinds of settings as my WSL2-related needs continue to evolve
No comments:
Post a Comment