LVM / local directory question

C

chrisalavoine

Renowned Member
Sep 30, 2009
152
0
81
Hi guys,

I have a single standalone Proxmox host that I wish to add to my existing cluster of 4 hosts - all running Proxmox 1.9.

The standalone host has a local directory called "VMDATA" which is just a local big disk.

The cluster uses an LVM iSCSI target called "VMDATA" which is connected to our SAN.

If I add the standalone server to the cluster without changing the storage name first will all hell break loose?

Kind Regards,
c:)
 
chrisalavoine said:
Is that in case there are any VMID conflicts?
Click to expand...

Storage configuration I also different (that was the initial question).

chrisalavoine said:
How about if I made sure that any existing VM's had different numbers from the existing cluster?
Click to expand...

Sure, you can manually fix all those things - if you can ;-)

Backup/restore is another way to do it (vzdump).
 
Hi Dietmar,

This is my /etc/pve/storage.cfg from each machine:

Standalone server:

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

dir: VMDATA
path /VMDATA
content images

dir: SNAPSHOTS
path /SNAPSHOTS
content backup


Cluster Master:

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

nfs: vm_snapshots2
path /mnt/pve/vm_snapshots2
server 192.168.x.x
export /c/vm_snapshots2
content backup

iscsi: VMDATA_target
portal 192.168.x.x
target iqn.2001-05.com.equallogic:0-8a0906-1a599c106-fea000000184bc81-vmdata
content none

lvm: VMDATA
vgname VMDATA
base VMDATA_target:0.0.0.scsi-36090a068109c591a81bc84010000a0fe
shared
content images

iscsi: ess-comms-001_target
portal 192.168.x.x
target iqn.2001-05.com.equallogic:0-8a0906-a7599c106-934000008f34d91f-ess-comms-001
content none

lvm: ess-comms-001
vgname ess-comms-001
base ess-comms-001_target:0.0.0.scsi-36090a068109c59a71fd9348f00004093
shared
content images

iscsi: ess-snapres_target
portal 192.168.x.x
target iqn.2001-05.com.equallogic:0-8a0906-e7099c106-e62000008fb4d926-ess-snapres
content none

lvm: ess-snapres
vgname ess-snapres
base ess-snapres_target:0.0.0.scsi-36090a068109c09e726d9b48f000020e6
shared
content images

iscsi: VMDATA2_target
portal 192.168.x.x
target iqn.2001-05.com.equallogic:0-8a0906-75456170a-32a922305084e1eb-misdata2
content none

lvm: VMDATA2
vgname VMDATA2
base VMDATA2_target:0.0.0.scsi-misDATA2
shared
content images


Can you see any conflicts there? I know that VMDATA exists on both servers but if one is a simple dir: entry and the other is lvm: surely they won't get in each others way? Or will they?
 
chrisalavoine said:
Can you see any conflicts there? I know that VMDATA exists on both servers but if one is a simple dir: entry and the other is lvm: surely they won't get in each others way? Or will they?
Click to expand...

VMDATA is a problem, because all VMs point to the wrong location after you copy the config from the master node.
 
Ok, thanks Dietmar.

I will snapshot all VM's then delete them, delete the VMDATA store, join the standalone server to the cluster and qmrestore all VM's back using the clusters SAN for storage.

c:)
 
You must log in or register to reply here.
Top Bottom