Ceph Config - Create FS - Error

[fxandrei]

So i am trying to configure proxmox to use ceph server (on 3 nodes\servers).
So i have created the osds, the pool, and the metadata servers.

So im following this : https://pve.proxmox.com/pve-docs/chapter-pveceph.html

So i added the ceph server to the cluster storage, and can use it.
Now im supposed to create the FS (even though i dont understant why exactly i need to do this... but ill keep investigating).

So if i try to create from the gui or from console (with "pveceph fs create --pg_num 128 --add-storage") i get this error :
creating data pool 'cephfs_data'...
mon_command failed - error parsing integer value '': Expected option value to be integer, got ''ta"}

If i check the pool after this i can see that it has create a new pool called "cephfs_data" with 128 pgnum.

So what can i do ?
I why would i need to create another pool with another pgnum when i already have a pool ?!

 

[Alwin]

So i added the ceph server to the cluster storage, and can use it.
Now im supposed to create the FS (even though i dont understant why exactly i need to do this... but ill keep investigating).

You only need to create a FS (and MDS) if you want to use CephFS. For VM disk storage, the RBD part is already sufficient.

So if i try to create from the gui or from console (with "pveceph fs create --pg_num 128 --add-storage") i get this error :
creating data pool 'cephfs_data'...
mon_command failed - error parsing integer value '': Expected option value to be integer, got ''ta"}

This seems like an already fixed bug, please upgrade to the latest packages.

 

[fxandrei]

What packages should have this fix ? And in what version ?

And is there a workaround ?
Could i create them manually somehow ?!

 

[Alwin]

What packages should have this fix ? And in what version ?

Best update the whole system.
https://pve.proxmox.com/pve-docs/chapter-sysadmin.html#_system_software_updates

And is there a workaround ?
Could i create them manually somehow ?!

With the ceph tooling.
http://docs.ceph.com/docs/luminous/cephfs/createfs/

 

[fxandrei]

So i just did the update :

apt-get dist-upgrade
Reading package lists... Done
Building dependency tree
Reading state information... Done
Calculating upgrade... Done
0 upgraded, 0 newly installed, 0 to remove and 0 not upgraded.

PS: i dont have enterprise support.
Is that a factor ?


I still get the same error.

Guess il try with the "ceph fs" command .

 

[fxandrei]

So something does not fit.
So when i get to the point of creating a CephFS i already have a pool created.
I cannot use that from what i can see.

I can create new pools, and the create the fs with these commands:

ceph osd pool create cephfs_data <pg_num>
ceph osd pool create cephfs_metadata <pg_num>
ceph fs new <fs_name> <metadata> <data>

If i execute ceph fs ls i see this :

root@hp1-s1:/var/lib# ceph fs ls
name: serv1-ceph-fs, metadata pool: serv1-ceph-metadata, data pools: [serv1-ceph-data ]
root@hp1-s1:/var/lib#

So whats the deal with the initial pool i created.
Should i delete it ?!
And what do i need now to do ?
Add the serv1-ceph-data to the storage cluster (RBD type).

What about the fs ?!

My question is what do i need to do to acces the pools !?

 

[fxandrei]

So yeah. Im kind of lost.
So i deleted everything.... so no pools.

I created the pools and fs manually with :

ceph osd pool create cephfs_data 512
ceph osd pool create cephfs_metadata 128

ceph fs new cephfs cephfs_metadata cephfs_data
ceph osd pool application enable cephfs_metadata cephfs

I then went to Storage in Proxmox Cluster and :
- added both clusterfs_data and clusterfs_metadata with RBD(PVE) storage type
- added clusterfs usint the clusterfs storage type

RBD can save disks, clusterfs cannot (just backups).

My problem is that im kind of lost .

Should i mount the metadata pool ? Or just the data ?!
How would i acces the disks i add in the data pool ? What about the metadata pool?

 

[fxandrei]

This one can be close, as i oppened another thread, only with my "problems" related to how to use these pools and fs.