I'm having trouble setting up some ceph osd's on my ceph cluster.
Currently I ave truenas scale for my nas server, but I want to change to a cep cluster, because aving block storage as a developer is kinda nice.
Currently i'm managing my cep osd via a raspberry pi5 wit an nvme extension board tat osts 2x 2TB samsung 970 evo's.
I have setup my cep cluster on multiple vm's on my vmware esxi.The problem i'm facing is that i can't create any osd, but I don't know why.
So my nvme's are: /dev/nvme0n1 and /dev/nvme1n1
when using the ui when going to -> physical disks, the nvme's are listed.
If i go to cluster->osds->create it does say no devices where found.When I click the button on the right bottom 'create osds' it keeps loading and does nothing.
So I tried via the command line using:ceph orch daemon add osd *<host>*:*<device-path>*where host = ceph-osd and device is /dev/nvme0n1 or /dev/nvme1n1
trying with the first I get gpt headers detected remove them first.using gdisk I removed the partitions on this drive, but it still does say that
- How do I remove tose headers?
when doing with the second nvme I get: 'Created no osd's on host ceph-osd. already created?
When checking on dashboard and cluster->osd tere are no osd's listed.
- What is te problem for the second drive?
I installed the cep cluster quiet recently so I should have the latest version from apt.All vm's are ubuntu 22.04.4
EDIT:
The gpt header issue is solved.
For the osd creation that something else.
I have added all my hosts and ip addresses in the host file of the cephadm.When I try todo a creation I get this error:
However it seems that the container is actually created, but it seems that it restarts and after this command failes when I try todo it again it tries to deploy a new container, since the names changes everytime.
On the cephadm log file in the ceph-osd host: it mentions : 'unable to create a new osd id'