Multi instances

Hi,
I am a newbie OpenNebula user and I have a problem…
I just finished installing the opennebula system with only one kvm virtualization host (centos 6.9 final) and I tried to run multi istances of ttylinux-kvm images but only the first start regularly the others remain in Pending status. Then I tried to run ttylinux-kvm and Ubuntu 16.04 - LXD together but again only the first runs… the second stay pending.
where is the error?
thank you all

@fbordini, you need to install LXDoNe addon in order to manage LXD containers on OpenNebula. Ubuntu 16.04 - LXD is an LXD container image.
OpenNebula doesn’t support multi instances nodes by default. As a workaround, you can follow the guides to prepare the same node to run KVM virtual machines and LXD containers and add it to OpenNebula twice (as a KVM node and as an LXD node). This way you can deploy KVM VMs and LXD containers on the same node :wink:

thanks, for information but with more instances of ttylinux only the first start normanly, the others stay panding. is it the same problem?

Hi!

Can you put here the sched.log file (/var/log/one/sched.log)?

Regards!
Juan Jose

sorry, but new users can not upload files…
Can I insert log in the answer?

Hi,

A link to pastebin or similar service should be ok too.

BR,
Anton Todorov

Hi,
this is the link with the last output log…
https://pastebin.com/GisjCXcZ
I don’t understand ‘there is no suitable host’ I have made two hosts but only can run KVM virtual machines.
Both instances must runs in same hosts…

I’ve never seen that in the logs

Have you tried to instantiate it manually? There’s no much info on the sched.log more than the scheduler couldn’t find a host that matches the criteria to deploy VMs :slightly_frowning_face:

I have not tried manually, but I have installed a new host node and I have instanced a second VM. It works … but I must create N identical VMs in a same node and now it does not work. The next week I will try manually … I keep you updated.
Bye

Hi,
I’ll try to execute manually three VMs and the result is …

onetemplate instantiate 1

VM ID: 51

onetemplate instantiate 1

VM ID: 52

onetemplate instantiate 1

VM ID: 53

onevm list

ID USER     GROUP    NAME            STAT UCPU    UMEM HOST             TIME
51 oneadmin oneadmin ttylinux - kvm- unkn    0      0K NebulaKVM    0d 00h01
52 oneadmin oneadmin ttylinux - kvm- unkn    0      0K HostMaster   0d 00h01
53 oneadmin oneadmin ttylinux - kvm- pend    0      0K              0d 00h00

the last instances does not start…

Hi!

Can you put the output of the onehost list command?

The scheduler can not find any active Host.

Regards!

ID NAME CLUSTER RVM ALLOCATED_CPU ALLOCATED_MEM STAT
5 HostMasterFews default 0 0 / 800 (0%) 0K / 15.6G (0%) on
7 NebulaKVM default 1 10 / 800 (1%) 1024M / 15.7G (6%) on

Now I have the following error in both nodes …

Mon Feb 5 16:05:31 2018 : Error monitoring Host HostMasterFews (5): …/…/vmm/kvm/poll:228:in round': wrong number of arguments (1 for 0) (ArgumentError) from ../../vmm/kvm/poll:228:inget_cpu_info’
from …/…/vmm/kvm/poll:226:in each' from ../../vmm/kvm/poll:226:inget_cpu_info’
from …/…/vmm/kvm/poll:132:in get_all_vm_info' from ../../vmm/kvm/../lib/poll_common.rb:99:inprint_all_vm_template’
from …/…/vmm/kvm/poll:531
ERROR MESSAGE --8<------
Error executing poll.sh
ERROR MESSAGE ------>8–
ARCH=x86_64
MODELNAME="Dual-Core AMD Opteron™ Processor 8212"
HYPERVISOR=kvm
TOTALCPU=800
CPUSPEED=2009
TOTALMEMORY=16332156
USEDMEMORY=397388
FREEMEMORY=15934768
FREECPU=800
USEDCPU=0
NETRX=30804822916
NETTX=778280921
DS_LOCATION_USED_MB=5943
DS_LOCATION_TOTAL_MB=50269
DS_LOCATION_FREE_MB=41767
DS = [
ID = 0,
USED_MB = 5943,
TOTAL_MB = 50269,
FREE_MB = 41767
]
VM=[ID=56,POLL="DISK_SIZE=[ID=0,SIZE=27] DISK_SIZE=[ID=1,SIZE=1] "]
HOSTNAME=HostMasterFews

… but the instances work fine…

I have create a my VM image … but on the same host the second one remains pending and the host status return the error above…I do not know what to do …:disappointed_relieved:

Have you followed the LXDoNe setup guide step by step?

Sorry but I do not understand … why do I have to install LXDoNe feature? I followed the main setup guide.
:thinking:

LXDoNe is an addon, its not provided by default by OpenNEbula. If you want OpenNebula to manage LXD containers then yo have to install the addon.

But also a personal Image made by iso needs of the the lxd addon?