PVE Cluster shows wrong storages?

[TechArea]

Hi

I've installed 4 old mac mini's and created a cluster.
Everything was ok, but something looks weird.
Why does I see the internal storage of device 1 on device 2 and device 3 and 4?
Also the available space is wrong, except on the host with physical connection to the disk.
For ex. bbb-disk2 has 500GB and is show right on pve-bbb but on pve-ccc/ddd/eee the disks has only 100GB

What's wrong here?

(take a look at the screenshot).

 

[Alwin]

You can see under Datacenter -> Storage all configured storage, there you can limit those for a specific host. Every host gets a local storage '/var/lib/vz', so those should be on every host locally available. There the shared flag is important.

 

[TechArea]

I'm not sure, does I understand you right?
The Disks aren't shared, except one.
But I see all the disks on all the pve's

 

[dcsapak]

when editing the storage, there is a 'nodes' field where you can select on which nodes the storage is available,
the shared flag is for marking storages as shared (as opposed to local) so that pve know it does not have to move/copy disks in case of migration etc.

 

[TechArea]

Now I get it!
Thank you!

But why the other pve's can't see the omv01 content? It's an nfs share which is attached at the "datacenter".
Take a look at the Screenshot

 

[dcsapak]

can you post the storage config ?

/etc/pve/storage.cfg

 

[TechArea]

dir: local
path /var/lib/vz
content vztmpl,backup,iso

lvmthin: local-lvm
thinpool data
vgname pve
content images,rootdir

dir: ssd01
path /media/ssd01
content images,rootdir,iso,vztmpl
nodes one
shared 0

dir: ssd02
path /media/ssd02
content backup
maxfiles 1
nodes one
shared 0

dir: storage01
path /media/storage01
content rootdir,iso,backup,vztmpl,images
maxfiles 1
nodes one
shared 0

nfs: omv01
export /export/omv01
path /mnt/pve/omv01
server 10.101.101.21
content backup,iso,vztmpl,rootdir,images
maxfiles 1
options vers=3

dir: bbb-disk2
path /media/bbb-disk2
content images,rootdir,iso,backup,vztmpl
maxfiles 1
nodes bbb
shared 0

 

[dcsapak]

is that ip reachable from all hosts? if yes what does 'pvesm status' say on the node where it does not work?

 

[TechArea]

Ok. the nfs seems to work, if I choose first an other host, select the nfs target and then switching the host... take a look at the gif.
Here is the output of the pvesm status (screenshot)

 

[dcsapak]

yeah i can see you have a storage 'asdf' defined which is not reachable on 'ccc' which blocks the api call (because it is trying to show all available storages, but this does not work)
the '?' in the tree is an indicator that that storage is not reachable

 

[TechArea]

Ok. I can't understand, why it works if I choose Host "one", because the asdf Storage is not properly mounted on any host. It was an unsuccessfull try to bind an iSCSI connection.

 

[Olaf Brenn]

Hi, I have to get back to that discussion. I have pretty much the same issue that a disk from PVE1 ist limited to 100 GB on PVE2. For backup strategy I would like to have full access from PVE to all disks. I tried the shared option with no results.
pvesm status shows all drives but with 100 GB limit only - on both PVE's .

 

[Olaf Brenn]

is there anybody with an Idea?

 

[Alwin]

@Olaf Brenn, please open up a new thread for your question it seems to diverge from OPs post and to get more visibility.

 

[Olaf Brenn]

moved to https://forum.proxmox.com/threads/pve-cluster-shows-wrong-storages.58545/