Could not create domain from / var / lib / one /datastores/0/9/deployment.3

(amira) #1

please is there anyone who can help me, i installed opennebula, on a ubuntu server frontend 16.04 and a kvm node on ubuntu server 16.04, when i deploy a virtual machine an error occurred Error deploying virtual machine: Could not create domain from / var / lib / one /datastores/0/9/deployment.3
ps I followed the verification procedure of the cloud installation in the site opennebula, it is started without any error

(swastik) #2

Hey @amira

Could you provide the complete log of the error from the log window of your VM instance?

(amira) #3

thank you for your reply, here is my log

(swastik) #4

Your disk.1 is not getting created on instantiating the VM, as it mentions no file or folder found. I am not sure how to rectify it. Can you check if the directory/file is present in the host or not, manually?

As VM images always have to be copied into the host so you need to make sure host is accessible by the front end server using ssh. (try if you can access host from front end server using ssh).

see if reading about the file system of opennebula helps…

(amira) #5

thanks, i’m using ssh transfert mode ,there is no problem with the connection between the host and the frontend i if i understand it’s problem with the host ???

(Amit Kumar) #6

Are you able to login compute host…Using one admin I’d… without password?

(amira) #7

thanks, yes, the connection between the host and the frontend without a password

(Михаил Знак) #8

i suppose that you have problem with datastores. i mean that you should check that system datastore has rigth id. i had the same and in my case it was incorrect system datastore

(Roland) #9

another thing to check: on the master server the user “oneadmin” has user ID 9869 and group ID 9869,
does the oneadmin user on the other server also have UID 9869 and GID 9869 ?

could you paste the output of:
onedatastore list
onedatastore show 0/1/2

(amira) #10

thanks, actually it have the same UID and GID in both master and host

(Михаил Знак) #11

is it actual problem? if ,yes, you can try edit /etc/libvirt/qemu.conf and change user/group for oneadmin. after restart libvirtd

(amira) #12

thanks, here is
onedatastore list
onedatastore show 0

onedatastore show 1

onedatastore show 2

(amira) #13

it is actually

(Михаил Знак) #14

try to add your own datastores (system, files, images) and deploy your vm on it

(amira) #15

i creates the datastores

and now it stucks on prolog for one hour , on the node u
and also in sched.log, i had this

(Михаил Знак) #16

Try to create your own cluster, and make for it all 3 datastores. after this try again. mode of datastore choose shared mode

(Roland) #17

there is something wrong with your system DS. This is the datastore where started VMs keep their files, in my system it shows the size and free space, in your screenshot size seems unknown.
This happens when opennebula cannot reach the datastore.
Also, the system datastore has set: SHARED=no, which prob. means that a hypervisor or host cannot access the system datastore?
If all is well, the TOTAL / FREE / USED / LIMIT values would have a value, it looks unaccessible now.

(amira) #18

thank you for your answer I am really blocked, I made the installation again and always the same problem, if I understood it is a problem of storage space for the ds, I will try to change the transfer mode from ssh to nfs

(amira) #19

finally, I solved this problem, it was because of the firewall that was active, I would like to thank you all for your feedbacks

(Roland) #20

good to hear you got passed the problem & thanks for posting the solution!