How to precisely check the actual disk usage of Ceph RBD?

[pulipulichen]

Hi everyone, I've run into a serious issue while managing a Proxmox VE Ceph environment.

A user created a lot of VMs and ended up filling the entire Ceph cluster. The problem is, when I look at the RBD storage in the WebUI, I can only see the "Provisioned Size" of each disk. I can't tell which VMs are actually writing large amounts of data.
To make matters worse, deleting some large disks didn't seem to free up space as expected. It feels like the storage is consumed and never released.

Is there a command or a way, similar to du on a regular filesystem, that allows me to see the "actual" disk usage so I can pinpoint exactly which VM is the space hog?

 

[aaron]

rbd -p <pool> du

would be the quickest way

 

[pulipulichen]

rbd -p <pool> du

would be the quickest way

You're absolutely right! That command worked perfectly, and I am able to successfully retrieve the actual disk usage. Thanks for the tip.

However, I have a follow-up question: Is there any way to have this information displayed directly in the Proxmox WebUI?

While the CLI is powerful, it’s not particularly convenient for regular users or for a quick glance at the status. Having this data visible in the GUI would be much more user-friendly for monitoring.

 

[aaron]

you can check if we already have a feature request in our bugtracke or not. if we do, please chime in
https://bugzilla.proxmox.com

without checking in detail, fetching that info would mean extending the storage plugin API, so a little bit more involved. But definitely make it known on the bugtracker, the more demand we see, the more likely we will work on it