Kshared Folder Top -

On the node, find processes accessing the mount:

lsof +D /var/lib/kubelet/pods/.../mount | grep -E "python|java|node|bin"

Then map process ID to pod:

cat /proc/<PID>/cgroup | grep kubepods

Or use kubectl debug node:

kubectl debug node/<node-name> -it --image=ubuntu
chroot /host
iotop -o -d 2

This shows real-time disk I/O per process, including NFS reads/writes.

Cause: Transient mount configured in /etc/fstab with the _netdev option but systemd ordering is wrong. Solution: Change your fstab line to: kshared_top /mnt/kshared virtiofs defaults,_netdev,x-systemd.after=virtiofsd.service 0 0 kshared folder top

| Element | Description | |---------|-------------| | Folder Icon | Standard shared folder emblem (overlay arrow / two people) | | Folder Name | Editable inline (if user has write perms on share metadata) | | Share Type Badge | SMB / NFS / FTP / WebDAV (color-coded) | | Access Status | “Shared with: Everyone” or “Only specific users” | | User Count | e.g., “👥 3 connected” (for active shares) | | Lock / Security | Shows encryption / password protection status | | Overflow Menu | ⋮ — contains: Modify share settings, Stop sharing, Copy share link, Show mount command |

Enable nfsiostat collector in node_exporter: On the node, find processes accessing the mount:

--collector.nfsiostat
--collector.nfsiostat.mount-points=/var/lib/kubelet/pods/.../mount

In a standard read() syscall, the CPU copies data from kernel memory to a user buffer.

Even the "top" setups fail. Here is how to fix the most common errors. Then map process ID to pod: cat /proc/&lt;PID&gt;/cgroup

The kshared architecture is built upon three foundational pillars: Memory Mapping, Access Control, and Synchronization.

Top