Hi Artur,

 

Apologies for the late response. Thanks for the below 😊

 

Yes, the 2 DC’s are located in two different physical locations. I will ask our data team if there is anything that they can see. This is a very strange issue. I see there was a similar issue that someone else also logged, but I see no solution has been posted. The subject of that email is: “VDSM HOST ISSUE - Message timeout which can be caused by communication issues”

 


Anton Louw
Cloud Engineer: Storage and Virtualization at Vox
T:  087 805 0000 | D: 087 805 1572
M: N/A
E: anton.louw@voxtelecom.co.za
A: Rutherford Estate, 1 Scott Street, Waverley, Johannesburg
www.vox.co.za

F
 
T
 
I
 
L
 
Y
 

From: Artur Socha <asocha@redhat.com>
Sent: 18 September 2020 14:45
To: Anton Louw <Anton.Louw@voxtelecom.co.za>
Cc: users@ovirt.org
Subject: Re: [ovirt-users] Re: Random hosts disconnects

 

 

 

On Fri, Sep 18, 2020 at 1:54 PM Anton Louw <Anton.Louw@voxtelecom.co.za> wrote:

 

Hi Artur,

 

Thanks for the reply. I have attached the system logs. There was a disconnect at 10:54, but no error that is different to the rest. I do see a whole lot of QEMU Guest Agent and block_io errors in the system logs. Not entirely sure what this means.

 

After a very quick search on the internet the first one does not seem to be severe at all - this guest agent provides only some information to VMs about the host.

Sep 18 10:50:41 node05.kvm.voxvm.co.za libvirtd[23603]: 2020-09-18 08:50:41.493+0000: 23729: error : qemuDomainAgentAvailable:9133 : Guest agent is not responding: QEMU guest agent is not connected

 

The second one is unknown to me at all:

ISep 18 10:50:52 node05.kvm.voxvm.co.za libvirtd[23603]: 2020-09-18 08:50:52.802+0000: 23729: error : qemuMonitorJSONBlockIoThrottleInfo:5005 : internal error: block_io_throttle inserted entry was not in expected format Sep 18

Perhaps someone with more libvirt/qemu background will comment on that.

 

 

Checking the vdsm logs at the time or the error, the only entry is the below:

 

“2020-09-18 10:55:57,081+0000 WARN  (qgapoller/2) [virt.periodic.VmDispatcher] could not run <function <lambda> at 0x7f2170395578> on ['d3838612-70bb-4731-a0d4-8f65d31b40a6', '59a2f394-48fe-4bd9-91d6-08115f2eec0a', 'f81e3ab8-c1a9-4674-b238-7e229fd43e7c', '42189fa1-4381-02c7-d830-20eac408da2c', '423f1c57-f98e-707f-c0f9-d4958d3f0fec', '64d1eabc-20ff-4288-98ff-dcfd120fe7d2', '4218baf0-e2a1-42c7-2efd-077407f47b4d', '42184650-5a60-5403-d758-840bdbf92dd8', '492ea3fe-0a27-4dde-abf9-7d146ee1b988', '4218df00-15cd-bdf9-efd9-c5ead49fd89c', '9c373379-718b-4906-abc1-960fb1820c2d', 'b9441c7a-0bfd-4d41-a8de-ee24e4259b36', 'd810325a-1a45-4054-a870-c8c052a22354', '42189d3f-4570-45ea-6e5a-94c85a5885a1'] (periodic:289)”

 

This WARN does not seem to be the cause ... it may be be the result because VM failed to be dispatched (perhaps due to lack of suitable hosts that got disconnected at a moment)

 

 

I am stumped. Do you think it is worth a shot increasing the vdsConnectionTimeout and vdsHeartbeatInSeconds to 40 for testing purposes?

 

I still don't think it will change anything unless your network between those 2 DC is 'tcp over pigeons' kind of setup :)

Now, more seriously. Even if increasing timeouts would fix the connectivity I suspect the core issue would still remain ... in the best case scenario it could be postponed a bit.

 

Am I correct assuming those 2 DC are  located in 2 different physical locations?   If so then I would closely check the network itself first (including hardware like routers/switches).

 

Artur

 

Thanks

 

 

Anton Louw

Cloud Engineer: Storage and Virtualization at Vox

T:  087 805 0000 | D: 087 805 1572
M: N/A
E: anton.louw@voxtelecom.co.za
A: Rutherford Estate, 1 Scott Street, Waverley, Johannesburg
www.vox.co.za

 

F

 

T

 

I

 

L

 

Y

 

 

From: Artur Socha <asocha@redhat.com>
Sent: 18 September 2020 13:27
To: Anton Louw <Anton.Louw@voxtelecom.co.za>
Cc: users@ovirt.org
Subject: Re: [ovirt-users] Re: Random hosts disconnects

 

Hi Anton,

I am not sure if changing this value would fix the issue. Defaults are pretty high. For example vdsHeartbeatInSeconds=30seconds, vdsTimeout=180seconds, vdsConnectionTimeout=20seconds.

 

Do you still have relevant logs from the affected hosts:

 /var/logs/vdsm/vdsm.log

 /var/logs/vdsm/supervdsm.log

Please look for any jsonrpc errors ie. write/read errors or (connection) timeouts.  Storage related warnings/errors might also be relevant.

 

Plus system logs if possible:

journalctl -f /usr/share/vdsm/vdsmd

journalctl  -f /usr/sbin/libvirtd

 

In order to get system logs from particular time period please combine it with the following example using -S  -U options:

journalctl -S "2020-01-12 07:00:00" -U "2020-01-12 07:15:00" 

I haven't a clue what to look there for besides any warnings/errors or anything else that seems .... unusual.

 

Artur

 

 

On Thu, Sep 17, 2020 at 8:09 AM Anton Louw via Users <users@ovirt.org> wrote:

 

Hi Everybody,

 

Did some digging around, and saw a few things regarding “vdsHeartbeatInSeconds”
I had a look at the properties file located at /etc/ovirt-engine/engine-config/engine-config.properties, and do not see an entry for “vdsHeartbeatInSeconds.type=Integer”. 
Seeing as these data centers are geographically split, could the “vdsHeartbeatInSeconds” potentially be the issue? Is it safe to increase this value after I add “vdsHeartbeatInSeconds.type=Integer” into my engine-config.properties file?
 
Thanks

 

 

Anton Louw

Cloud Engineer: Storage and Virtualization at Vox

T:  087 805 0000 | D: 087 805 1572
M: N/A
E: anton.louw@voxtelecom.co.za
A: Rutherford Estate, 1 Scott Street, Waverley, Johannesburg
www.vox.co.za

 

F

 

T

 

I

 

L

 

Y

 

 

From: Anton Louw via Users <users@ovirt.org>
Sent: 16 September 2020 09:01
To: users@ovirt.org
Subject: [ovirt-users] Random hosts disconnects

 

 

Hi All,

 

I have a strange issue in my oVirt environment. I currently have a standalone manager which is running in VMware. In my oVirt environment, I have two Data Centers. The manager is currently sitting on the same subnet as DC1. Randomly, hosts in DC2 will say “Not Responding” and then 2 seconds later, the hosts will activate again.

 

The strange thing is, when the manager was sitting on the same subnet as DC2, hosts in DC1 will randomly say “Not Responding”

 

I have tried going through the logs, but I cannot see anything out of the ordinary regarding why the hosts would drop connection. I have attached the engine.log for anybody that would like to do a spot check.

 

Thanks

 

Anton Louw

Cloud Engineer: Storage and Virtualization at Vox

T:  087 805 0000 | D: 087 805 1572
M: N/A
E: anton.louw@voxtelecom.co.za
A: Rutherford Estate, 1 Scott Street, Waverley, Johannesburg
www.vox.co.za

 

F

 

T

 

I

 

L

 

Y

 

 

#VoxBrand


Disclaimer

The contents of this email are confidential to the sender and the intended recipient. Unless the contents are clearly and entirely of a personal nature, they are subject to copyright in favour of the holding company of the Vox group of companies. Any recipient who receives this email in error should immediately report the error to the sender and permanently delete this email from all storage devices.

This email has been scanned for viruses and malware, and may have been automatically archived by Mimecast Ltd, an innovator in Software as a Service (SaaS) for business. Providing a safer and more useful place for your human generated data. Specializing in; Security, archiving and compliance. To find out more Click Here.

 

 

_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/privacy-policy.html
oVirt Code of Conduct: https://www.ovirt.org/community/about/community-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@ovirt.org/message/EJL246IPBGEHIQ5KUWG2APSTQWFE7VFK/



--

Artur Socha
Senior Software Engineer, RHV
Red Hat

 



--

Artur Socha
Senior Software Engineer, RHV
Red Hat