[ovirt-users] Unable to install 3.6.5 under CentOS 6.7

[Tom Lambdin]

Hi,

Although the release note indicates that 3.6.5 is supported under CentOS/RHEL 
6.7, when we attempt to update to the the latest version under CentOS 6.7, the 
highest version we can get from the repo is 3.6.3.
If we spin up a brand new machine running CentOS 7.2, we do however 
successfully get 3.6.5. Ideally we’d like to avoid having to migrate our 
management host to CentOS 7 at this time.
Is there something we’re missing here or is there a problem with the repo 
configuration?

Regards,

-Tom Lambdin



signature.asc
Description: Message signed with OpenPGP using GPGMail
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Unable to install 3.6.5 under CentOS 6.7

[Yedidyah Bar David]

On Mon, May 16, 2016 at 12:58 PM, Tom Lambdin  wrote:
> Hi,
>
> Although the release note indicates that 3.6.5 is supported under CentOS/RHEL 
> 6.7, when we attempt to update to the the latest version under CentOS 6.7, 
> the highest version we can get from the repo is 3.6.3.
> If we spin up a brand new machine running CentOS 7.2, we do however 
> successfully get 3.6.5. Ideally we’d like to avoid having to migrate our 
> management host to CentOS 7 at this time.
> Is there something we’re missing here or is there a problem with the repo 
> configuration?

The repo does contain ovirt-engine-3.6.5.3-1.el6.noarch.rpm [1].

Did you try 'yum clean all'?

Please post your repo files and output of relevant yum commands.

Thanks,

[1] 
http://plain.resources.ovirt.org/pub/ovirt-3.6/rpm/el6/noarch/ovirt-engine-3.6.5.3-1.el6.noarch.rpm

>
> Regards,
>
> -Tom Lambdin
>
>
> ___
> Users mailing list
> Users@ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>



-- 
Didi
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Unable to install 3.6.5 under CentOS 6.7

[Tom Lambdin]

Hi,

So I’ve just checked again and it now works. Very strange. Thanks for the quick 
response anyway.

-Tom

> On 16 May 2016, at 11:12, Yedidyah Bar David  wrote:
> 
> On Mon, May 16, 2016 at 12:58 PM, Tom Lambdin  wrote:
>> Hi,
>> 
>> Although the release note indicates that 3.6.5 is supported under 
>> CentOS/RHEL 6.7, when we attempt to update to the the latest version under 
>> CentOS 6.7, the highest version we can get from the repo is 3.6.3.
>> If we spin up a brand new machine running CentOS 7.2, we do however 
>> successfully get 3.6.5. Ideally we’d like to avoid having to migrate our 
>> management host to CentOS 7 at this time.
>> Is there something we’re missing here or is there a problem with the repo 
>> configuration?
> 
> The repo does contain ovirt-engine-3.6.5.3-1.el6.noarch.rpm [1].
> 
> Did you try 'yum clean all'?
> 
> Please post your repo files and output of relevant yum commands.
> 
> Thanks,
> 
> [1] 
> http://plain.resources.ovirt.org/pub/ovirt-3.6/rpm/el6/noarch/ovirt-engine-3.6.5.3-1.el6.noarch.rpm
> 
>> 
>> Regards,
>> 
>> -Tom Lambdin
>> 
>> 
>> ___
>> Users mailing list
>> Users@ovirt.org
>> http://lists.ovirt.org/mailman/listinfo/users
>> 
> 
> 
> 
> --
> Didi



signature.asc
Description: Message signed with OpenPGP using GPGMail
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

[ovirt-users] How to migrate un-migratable VMs?

[Nicolas Ecarnot]

Hi,

Some VMs are doing high guest-memory writes (mainly files and mail 
servers, but not only), so migrations are reproductively failing.

I found some info here :
https://bugzilla.redhat.com/show_bug.cgi?id=919201
that confirm what I was fearing.

This BZ is aging from dinosaurs times, so I was wondering now we're in 
the 3.6.x age and 4.x coming, are there advices one may give to migrate 
those verbose VMs?


I guess increasing the allowed page memory is not a correct answer as it 
will fill as before.


Any thoughts?

--
Nicolas ECARNOT
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

[ovirt-users] disk order in vm and disk order in export

[Fabrice Bacchella]

I have a VM with the following disks (in the ordrer return by a get query):
 

...
vm_sys
17179869184 <- 16GiB
virtio_scsi
true
image



vm_data
virtio_scsi
false


1099511627776 <- 1 TiB.
0


lun


So the first one is a 16 GiB virtual disk, the second one a 1 iTB direct access 
LUN that will contains data.

But once booterd, I see:
lsblk   
NAMEMAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda   8:00 1T  0 disk 
...
sdb   8:16   016G  0 disk 
...

and :
ls -l /dev/disk/by-*/*
...
lrwxrwxrwx. 1 root root  9 May 16 09:03 
/dev/disk/by-path/pci-:00:04.0-scsi-0:0:0:0 -> ../../sda
lrwxrwxrwx. 1 root root  9 May 16 09:03 
/dev/disk/by-path/pci-:00:04.0-scsi-0:0:0:1 -> ../../sdb

So the first disk sda is the second non bootable disk.
What set the disk order associated to a VM ? Is there a way to enforce that the 
first disk is really the first disk ?___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

[ovirt-users] oVirt Node 3.6

[Anantha Raghava]

Hi,

I attempted to download the oVirt Node 3.6 from 
http://resources.ovirt.org/pub/ovirt-3.6-snapshot/iso/ovirt-node-iso/ 
and I also find there is another ovirt-node-ng. Which one should we 
download.


I downloaded the one at 
http://resources.ovirt.org/pub/ovirt-3.6-snapshot/iso/ovirt-node-iso/, 
but this iso file seems to have an issue. Although it boots the server 
properly at around 66% into installation, the installer throws up an 
error - I/O error, file not found or some thing of this sort. 
Installation doesn't go forward at all. We have to power off the server, 
reboot with CentOS DVD install using yum local install command.


Are these files proper? Which one should we use?
--

Thanks & Regards,


Anantha Raghava



___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] oVirt & gluster - which copy is written to?

[Alexander Wels]

On Thursday, May 12, 2016 12:05:22 AM Bill Bill wrote:
> Hello,
> 
> Let’s say I have a 5 node converged cluster of oVirt & glusterFS with a
> replica count of “3”.
 
> Host1 - replica
> Host2 - replica
> Host3 - replica
> Host4
> Host5
> 
> If I spin a VM up on Host1 – does the first replica get created local to
> that server?
 
> In the event I move the VM to another host, does the replica of the disk
> follow the VM so essentially it’s writing to a local copy first or does it
> “always” write across the network.
 
> For example, with VMware, the disk will follow the VM to the hypervisor you
> move it to so that the writes are essentially “local”.
 
> I have not been able to find anything on the internet or in the
> documentation that specifies how this process works. Ideally, you wouldn’t
> want a VM residing on Host1 that accesses its disk on Host3. Multiple that
> by hundreds of VM’s and it would create a ton of unnecessary network
> congestion.
 
While I cannot answer your actual question as I don't know enough about the 
internal implementation of how oVirt interacts with gluster. I just want to 
point out major difference between oVirt and VMWare. In oVirt storage domains 
are attached to data centers, not hosts. So there is no does the disk move 
from host1 to host2. The disk is on the storage domain and host1 and host2 and 
every other host in the data center has access to the storage domain.


___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] USB

[Brett I. Holcomb]

On Sun, 2016-05-15 at 14:40 +0300, Nisim Simsolo wrote:
> Hi
> 
> You can attach host device directly from virtual machines tab ->
> select VM -> host devices sub tab -> "add device" button.
> 
> But first you need to verify your BIOS is supporting intel VT-d or
> AMD vi.
> And you also need to activate it in the kernel. The next link
> describes how to do it: 
> https://access.redhat.com/documentation/en-US/Red_Hat_Enterprise_Linu
> x/7/html/Virtualization_Deployment_and_Administration_Guide/chap-
> Guest_virtual_machine_device_configuration.html#sect-
> Guest_virtual_machine_device_configuration-PCI_devices
> 
> note: Instead of adding "intel_iommu=pt" in the grub cmdline, add
> "intel_iommu=on" or "amd_iommu=on".
> 
> Thank you  Nsim and Martin.  I'll give that  a try.
As a point of information has anyone tried USB/IP or Incentives Pro?___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] Adding another host to my cluster

[Gervais de Montbrun]

Hi Roy,

I've added /var/log/ovirt-engine/engine.log to my DropBox: 
https://www.dropbox.com/sh/21lkl6ao3hv2r85/AADDMCK4Dp7UYgs5P6dMQxHKa/vdsm?dl=0 


When I look in the engine.log, I see the error "unable to load number from 
serial.txt"
I did a find for serial.txt and the only file that comes up is: 
/etc/pki/ovirt-engine/serial.txt 
(Makes sense that this is the file it wants to read)
When I open /etc/pki/ovirt-engine/serial.txt, it is an empty file. 
[root@cultivar ovirt-engine]# cat /etc/pki/ovirt-engine/serial.txt
[root@cultivar ovirt-engine]# 

What should be in serial.txt?

I don't understand where to find the other log that you mention in your email 
below. I don't see any paths listed in the UI when I click on the failed event 
for certificate enrollment.

Cheers,
Gervais



> On May 15, 2016, at 5:10 PM, Roy Golan  > wrote:
> 
> 
> 
> On Sun, May 15, 2016 at 1:32 AM, Gervais de Montbrun  > wrote:
> Hey Folks,
> 
> I'm still stuck here... I believe the issue may be because the Enrolling 
> certificates stage is failing...
> 
> Any idea's how to sort this out?
> 
> I see:
>   Installing Host Cultivar2. Enrolling certificate
>   Failed to install Host Cultivar2. Certificate enrollment failed.
>   Host Cultivar2 installation failed. Certificate enrollement failed.
> 
> Any attempt to re-install fails at this step.
> 
> Cheers,
> Gervais
> 
> 
> 
> I also see in the logs that the manager (ovirt-engine) rejected the host 
> addition to the setup. Please add the /var/log/ovirt-engine/engine.log from 
> the engine machine. Also, the engine machine has the install log - open one 
> of the Event logs in the UI that says you host installation failed and you 
> will see it retrieved the install log to a path under your engine. Please 
> include this as well.
>  
> 
>> On May 13, 2016, at 1:36 PM, Gervais de Montbrun > > wrote:
>> 
>> Hi Nir,
>> 
>> Thank you for you input.
>> 
>> It was suggest that I try starting vdsm from the shell so I could see the 
>> output in an effort to solve my issue: 
>> http://lists.ovirt.org/pipermail/users/2016-May/039690.html 
>> 
>> 
>> Cheers,
>> Gervais
>> 
>> 
>> 
>>> On May 13, 2016, at 12:12 PM, Nir Soffer >> > wrote:
>>> 
>>> On Fri, May 13, 2016 at 3:37 PM, Gervais de Montbrun
>>> mailto:gerv...@demontbrun.com>> wrote:
 Hi Charles,
 
 I think the problem I am having is due to the setup failing and not
 something in vdsm configs as I have never gotten this server to start up
 properly and the BRIDGE ethernet interface + ovirt routes are not setup.
 
 I put the logs here:
 https://www.dropbox.com/sh/5ugyykqh1lgru9l/AACXxRYWr3tgd0WbBVFW5twHa?dl=0 
 
 
 hosted-engine--deploy-logs.zip # Logs from when I tried to deploy and it
 failed
 vdsm.tar.gz # /var/log/vdsm
 
 Output from running vdsm from the command line:
 
 [root@cultivar2 log]# su -s /bin/bash vdsm
>>> 
>>> This cannot work unless supervdsmd is running...
>>> 
 [vdsm@cultivar2 log]$ python /usr/share/vdsm/vdsm
 (PID: 6521) I am the actual vdsm 4.17.26-1.el7
 cultivar2.grove.silverorange.com 
  (3.10.0-327.el7.x86_64)
 VDSM will run with cpu affinity: frozenset([1])
 /usr/bin/taskset --all-tasks --pid --cpu-list 1 6521 (cwd None)
 SUCCESS:  = '';  = 0
 Starting scheduler vdsm.Scheduler
 started
 Run and protect:
 registerDomainStateChangeCallback(callbackFunc=>>> 0x381b158>)
 Run and protect: registerDomainStateChangeCallback, Return response: None
 Trying to connect to Super Vdsm
 Preparing MOM interface
 Using named unix socket /var/run/vdsm/mom-vdsm.sock
 Unregistering all secrests
 trying to connect libvirt
 recovery: started
 Setting channels' timeout to 30 seconds.
 Starting VM channels listener thread.
 Listening at 0.0.0.0:54321 
 Adding detector 
 recovery: completed in 0s
 Adding detector >>> 0x382e5a8>
 Starting executor
 Starting worker jsonrpc.Executor/0
 Worker started
 Starting worker jsonrpc.Executor/1
 Worker started
 Starting worker jsonrpc.Executor/2
 Worker started
 Starting worker jsonrpc.Executor/3
 Worker started
 Starting worker jsonrpc.Executor/4
 Worker started
 Starting worker jsonrpc.Executor/5
 Worker started
 Starting worker jsonrpc.Executor/6
 Worker started
 Starting worker jsonrpc.Executor/7
 Worker started
 XMLRPC server running
 Starting executor
 Starting worker periodic/0
 Worker started
 Starting worker periodic/1

Re: [ovirt-users] Adding another host to my cluster

[Gervais de Montbrun]

Roy,

Thank you so much for asking for the logs on the engine. The serial.txt bein 
blank was my issue. I found a solution in the following thread online here: 
http://lists.ovirt.org/pipermail/users/2014-November/028870.html
I did the exact same thing that Dinuwan did to solve his issue back in 2014 and 
was able to enrol my new server and install the hosted-engine just fine on it.

I am not certain what caused serial.txt to be empty and am will to share logs 
to anyone from the oVirt team who might want to see them. For my purposes, I'm 
considering this issue solved.

Thanks to everyone who offered advice and suggestions.

Cheers,
Gervais



> On May 16, 2016, at 2:04 PM, Gervais de Montbrun  
> wrote:
> 
> Hi Roy,
> 
> I've added /var/log/ovirt-engine/engine.log to my DropBox: 
> https://www.dropbox.com/sh/21lkl6ao3hv2r85/AADDMCK4Dp7UYgs5P6dMQxHKa/vdsm?dl=0
>  
> 
> 
> When I look in the engine.log, I see the error "unable to load number from 
> serial.txt"
> I did a find for serial.txt and the only file that comes up is: 
> /etc/pki/ovirt-engine/serial.txt 
> (Makes sense that this is the file it wants to read)
> When I open /etc/pki/ovirt-engine/serial.txt, it is an empty file. 
>   [root@cultivar ovirt-engine]# cat /etc/pki/ovirt-engine/serial.txt
>   [root@cultivar ovirt-engine]# 
> 
> What should be in serial.txt?
> 
> I don't understand where to find the other log that you mention in your email 
> below. I don't see any paths listed in the UI when I click on the failed 
> event for certificate enrollment.
> 
> Cheers,
> Gervais
> 
> 
> 
>> On May 15, 2016, at 5:10 PM, Roy Golan > > wrote:
>> 
>> 
>> 
>> On Sun, May 15, 2016 at 1:32 AM, Gervais de Montbrun > > wrote:
>> Hey Folks,
>> 
>> I'm still stuck here... I believe the issue may be because the Enrolling 
>> certificates stage is failing...
>> 
>> Any idea's how to sort this out?
>> 
>> I see:
>>  Installing Host Cultivar2. Enrolling certificate
>>  Failed to install Host Cultivar2. Certificate enrollment failed.
>>  Host Cultivar2 installation failed. Certificate enrollement failed.
>> 
>> Any attempt to re-install fails at this step.
>> 
>> Cheers,
>> Gervais
>> 
>> 
>> 
>> I also see in the logs that the manager (ovirt-engine) rejected the host 
>> addition to the setup. Please add the /var/log/ovirt-engine/engine.log from 
>> the engine machine. Also, the engine machine has the install log - open one 
>> of the Event logs in the UI that says you host installation failed and you 
>> will see it retrieved the install log to a path under your engine. Please 
>> include this as well.
>>  
>> 
>>> On May 13, 2016, at 1:36 PM, Gervais de Montbrun >> > wrote:
>>> 
>>> Hi Nir,
>>> 
>>> Thank you for you input.
>>> 
>>> It was suggest that I try starting vdsm from the shell so I could see the 
>>> output in an effort to solve my issue: 
>>> http://lists.ovirt.org/pipermail/users/2016-May/039690.html 
>>> 
>>> 
>>> Cheers,
>>> Gervais
>>> 
>>> 
>>> 
 On May 13, 2016, at 12:12 PM, Nir Soffer >>> > wrote:
 
 On Fri, May 13, 2016 at 3:37 PM, Gervais de Montbrun
 mailto:gerv...@demontbrun.com>> wrote:
> Hi Charles,
> 
> I think the problem I am having is due to the setup failing and not
> something in vdsm configs as I have never gotten this server to start up
> properly and the BRIDGE ethernet interface + ovirt routes are not setup.
> 
> I put the logs here:
> https://www.dropbox.com/sh/5ugyykqh1lgru9l/AACXxRYWr3tgd0WbBVFW5twHa?dl=0 
> 
> 
> hosted-engine--deploy-logs.zip # Logs from when I tried to deploy and it
> failed
> vdsm.tar.gz # /var/log/vdsm
> 
> Output from running vdsm from the command line:
> 
> [root@cultivar2 log]# su -s /bin/bash vdsm
 
 This cannot work unless supervdsmd is running...
 
> [vdsm@cultivar2 log]$ python /usr/share/vdsm/vdsm
> (PID: 6521) I am the actual vdsm 4.17.26-1.el7
> cultivar2.grove.silverorange.com 
>  (3.10.0-327.el7.x86_64)
> VDSM will run with cpu affinity: frozenset([1])
> /usr/bin/taskset --all-tasks --pid --cpu-list 1 6521 (cwd None)
> SUCCESS:  = '';  = 0
> Starting scheduler vdsm.Scheduler
> started
> Run and protect:
> registerDomainStateChangeCallback(callbackFunc= at
> 0x381b158>)
> Run and protect: registerDomainStateChangeCallback, Return response: None
> Trying to connect to Super Vdsm
> Preparing MOM interface
> Using named unix socket /var/run/vdsm/mom-vdsm.sock
> Unregistering all secrests
> trying to connect libvirt
> recovery: started
>

[ovirt-users] During Hosted Engine Deployment. CPU Type not listed as compatible. No selection to choose from.

[Langley, Robert]

Host is a Dell PowerEdge 2950 with an Intel Xeon 5160 (Woodcrest) processor.
I'm stuck at the prompt, "Please specify the CPU type to be used by the VM []" 
during the deployment.
There is nothing to select. No compatible CPU type is shown.
The server is from 2008 and I'm needing to use these older servers to get 
started, before I can bring in the newer ones that are busy hosting VMs.
The other Dell PE 2950 is from 2007 and I was able to bring up the engine on 
that one. It is a Conroe CPU. I won't get into my mistake with that deployment. 
I'm working on recovering from that one.
I wanted to post about this, because I was trying to add this one (2008) as an 
additional host and it wouldn't add, due to the processor.
So, one question (in addition to the problem above) is whether or not I can use 
two hosts with two different processors for the hosted engine?
Thanks
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users

Re: [ovirt-users] VMs seems to be slow, while snapshot delete is happening

[SATHEESARAN]

On 05/14/2016 02:38 AM, Greg Padgett wrote:

On 05/12/2016 03:46 AM, SATHEESARAN wrote:

Hi All,

I have created a VM with 60GB of disk space.
I have started I/O inside the VM, after sometime when the disk was 
4GB full,

I took a snapshot of the disk from UI.

I/O kept continuing and the new image file ( overlay file ) was 15GB in
size.

Now I deleted the snapshot.
I think this would initiate block-merges.
During this time, my I/O was were slower, till the snapshot delete is
completed.

Its not very slow, but slow.
Is this a known behavior ?


Hi Satheesaran,

It sounds like this was a live merge, ie snapshot removal while the VM 
was running.  This is currently implemented as a block commit, where 
data from the snapshot (your 15GB overlay, in this case) is merged down.


For snapshot deletion to complete, this means that your concurrent I/O 
and the data already in the overlay have to be written into the 
next-lower image.  So in short, yes, depending on the storage system 
backing your VM, I/O may slow down a bit until completion.


HTH,
Greg

Greg,

Thanks for that information.
I too expected it, but really wanted to confirm the behavior.

-- Satheesaran S
___
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users