I have been trying to install a Ceph MDS on Proxmox and am having difficulties.
First: I am aware that mds is not yet supported on proxmox.
Second: I have seen in the forum that it is likely possible.
Third: I want to look into it.
So I currently have configured a cluster with 3 Proxmox 3.3 with Ceph firefly nodes. All latest updates done on the no-subscription apt.
The Proxmox LAN is 1GB 192.168.255.0/24 and the storage LAN is a 2 X GB NICs bonded together using round-robin on a separate switch at 10.10.10.0/24.
The Proxmox/ceph nodes each have 1 X SSD for the system and 4 x 1TB WD HDD Red drives for OSDs.
I have connectivity everywhere and the Ceph health is OK.
So far, I have 2 KVM using RBD with acceptable performance.
I installed ceph-deploy on all three nodes in order to install a MDS:
Then, I tried to install one MDS:
I got:
I changed directory to where the ceph.conf is on Proxmox:
and now I am getting:
Running gatherkeys, I then get:
The link from /etc/ceph/ceph.conf -> /etc/pve/ceph.conf works but appears to cause some issue.
Can anyone shed some light on what I may be doing wrong?
This should not be this hard.
Please advise.
Serge.
First: I am aware that mds is not yet supported on proxmox.
Second: I have seen in the forum that it is likely possible.
Third: I want to look into it.
So I currently have configured a cluster with 3 Proxmox 3.3 with Ceph firefly nodes. All latest updates done on the no-subscription apt.
The Proxmox LAN is 1GB 192.168.255.0/24 and the storage LAN is a 2 X GB NICs bonded together using round-robin on a separate switch at 10.10.10.0/24.
The Proxmox/ceph nodes each have 1 X SSD for the system and 4 x 1TB WD HDD Red drives for OSDs.
I have connectivity everywhere and the Ceph health is OK.
So far, I have 2 KVM using RBD with acceptable performance.
I installed ceph-deploy on all three nodes in order to install a MDS:
Code:
apt-get install ceph-deploy
Then, I tried to install one MDS:
Code:
ceph-deploy mds create pve1
I got:
Code:
root@pve1:/etc/pve/priv/ceph# ceph-deploy mds create pve1
[ceph_deploy.conf][DEBUG ] found configuration file at: /root/.cephdeploy.conf
[ceph_deploy.cli][INFO ] Invoked (1.5.17): /usr/bin/ceph-deploy mds create pve1
[ceph_deploy][ERROR ] ConfigError: Cannot load config: [Errno 2] No such file or directory: 'ceph.conf'; has `ceph-deploy new` been run in this directory?
I changed directory to where the ceph.conf is on Proxmox:
Code:
cd /etc/pve
and now I am getting:
Code:
root@pve1:/etc/pve# ceph-deploy mds create pve1
[ceph_deploy.conf][DEBUG ] found configuration file at: /root/.cephdeploy.conf
[ceph_deploy.cli][INFO ] Invoked (1.5.17): /usr/bin/ceph-deploy mds create pve1
[ceph_deploy.mds][DEBUG ] Deploying mds, cluster ceph hosts pve1:pve1
[ceph_deploy][ERROR ] RuntimeError: bootstrap-mds keyring not found; run 'gatherkeys'
Running gatherkeys, I then get:
Code:
root@pve1:/etc/pve# ceph-deploy gatherkeys pve1
[ceph_deploy.conf][DEBUG ] found configuration file at: /root/.cephdeploy.conf
[ceph_deploy.cli][INFO ] Invoked (1.5.17): /usr/bin/ceph-deploy gatherkeys pve1
[ceph_deploy.gatherkeys][DEBUG ] Checking pve1 for /etc/ceph/ceph.client.admin.keyring
[pve1][DEBUG ] connected to host: pve1
[pve1][DEBUG ] detect platform information from remote host
[pve1][DEBUG ] detect machine type
[pve1][DEBUG ] fetch remote file
[ceph_deploy.gatherkeys][DEBUG ] Got ceph.client.admin.keyring key from pve1.
[ceph_deploy.gatherkeys][DEBUG ] Checking pve1 for /var/lib/ceph/mon/ceph-{hostname}/keyring
[pve1][DEBUG ] connected to host: pve1
[pve1][DEBUG ] detect platform information from remote host
[pve1][DEBUG ] detect machine type
[pve1][DEBUG ] fetch remote file
[ceph_deploy.gatherkeys][WARNIN] Unable to find /var/lib/ceph/mon/ceph-{hostname}/keyring on ['pve1']
[ceph_deploy.gatherkeys][DEBUG ] Checking pve1 for /var/lib/ceph/bootstrap-osd/ceph.keyring
[pve1][DEBUG ] connected to host: pve1
[pve1][DEBUG ] detect platform information from remote host
[pve1][DEBUG ] detect machine type
[pve1][DEBUG ] fetch remote file
[ceph_deploy.gatherkeys][DEBUG ] Got ceph.bootstrap-osd.keyring key from pve1.
[ceph_deploy.gatherkeys][DEBUG ] Checking pve1 for /var/lib/ceph/bootstrap-mds/ceph.keyring
[pve1][DEBUG ] connected to host: pve1
[pve1][DEBUG ] detect platform information from remote host
[pve1][DEBUG ] detect machine type
[pve1][DEBUG ] fetch remote file
[ceph_deploy.gatherkeys][DEBUG ] Got ceph.bootstrap-mds.keyring key from pve1.
root@pve3:/etc/pve# ceph-deploy mds create pve1
[ceph_deploy.conf][DEBUG ] found configuration file at: /root/.cephdeploy.conf
[ceph_deploy.cli][INFO ] Invoked (1.5.17): /usr/bin/ceph-deploy mds create pve1
[ceph_deploy.mds][DEBUG ] Deploying mds, cluster ceph hosts pve1:pve1
[pve1][DEBUG ] connected to host: pve1
[pve1][DEBUG ] detect platform information from remote host
[pve1][DEBUG ] detect machine type
[ceph_deploy.mds][INFO ] Distro info: debian 7.6 wheezy
[ceph_deploy.mds][DEBUG ] remote host will use sysvinit
[ceph_deploy.mds][DEBUG ] deploying mds bootstrap to pve1
[pve1][DEBUG ] write cluster configuration to /etc/ceph/{cluster}.conf
[ceph_deploy.mds][ERROR ] RuntimeError: config file /etc/ceph/ceph.conf exists with different content; use --overwrite-conf to overwrite
[ceph_deploy][ERROR ] GenericError: Failed to create 1 MDSs
The link from /etc/ceph/ceph.conf -> /etc/pve/ceph.conf works but appears to cause some issue.
Can anyone shed some light on what I may be doing wrong?
This should not be this hard.
Please advise.
Serge.