Question: Dashboard counts available storage double ?

[Glowsome]

In my setup i have a 4-node cluster with shared SAS-Storage configured with Multipath.
When i look at the Datacenter-> Summary report i see alot more storage which is actually available being calculated.

As the usage of 'local' storage (in any way) within a cluster when having shared storage available is unwanted (as i see it).
On the Datacenter -> Storage view i see one 'local thing, which is disabled.
So in my assumption it should not be taken into account.

However if i were to calculate all this available (local) storage into the number i am seeing its still short.
My only idea as to explain this is that the Summary is calculated over (node-)available disks ( seen from a/the Disk-view on a node before being multipath'ed)


This would make sense, as prior to being multipath'ed - with 2 connections per disk would roughly sum up the toral storage i am seeing.

[edit] If i look at the LVM-view i see this :

Is anyone else experiencing this, and if so .. any solve ?

- Glowsome

 

[mira]

Could you elaborate a bit more what you'd expect to see?

The `Disk View` is always local to the node you're currently viewing.

 

[Glowsome]

@m

Could you elaborate a bit more what you'd expect to see?

The `Disk View` is always local to the node you're currently viewing.

Expectation is to see the correct amount of storage being offered to PVE instead of the now (seemingly) doubled amount displayed.
If you look at the LVM-screenie/view and calculate all space together you get to about 22TB , but was seeing 44TB total displayed on the Dashboard.

However somewhere it seems to have 'corected' itself .. i have done nothing more over the course of days (since creating the post) update the machines with packages ... and thus i can no longer even investigate the presented issue.

 

[mira]

Could it be that multipath wasn't configured correctly at that time and disks were seen as standalone, which would explain the doubled amount?

 

[Glowsome]

Could it be that multipath wasn't configured correctly at that time and disks were seen as standalone, which would explain the doubled amount?

When Multipath was introduced (with new servers rotated in) it was configured correctly since start.
On the old servers no Multipath was present and servers only had 1 connection to the storage.
I distinctly remember also seeing same behaviour back then, but as said since i created the post and performed regular package-updates ( to keep in check with latest) the issue resolved itself.