----- Original Message -----

Dne 28.11.2013 10:12, Mike Kolesnik napsal(a):

...

----- Original Message -----

...

Hello, Hi Jakub,

...

I found strange problem in vdsm 4.13.0 with oVirt 3.3.1. After 30 hours from update (ovirt -> 3.3.1, vdsm -> 4.13) my one vlan network does not work. There is nothing in log, I can see network trafic from not working vlan on node in tcpdump, but it not come to VM. Did it stop working all of the sudden or did you change something?

...

After vdsm restart on node it works like expected. Seems like VDSM restored configuration to last known config on that host, maybe there's something in vdsm.log to indicate what went wrong.

...

I take a look in to oVirt -> Host -> Network Interface and all nodes has bond interface changed from "Mode 4 Dynamic link agregation" to "Custom: mode=4" and Custom mode is filled by "mode=4" Mabe it is the same, but who change that? This is a know behavior but it's still mode 4 it's just the way it displays it.

...

I can have one node in this status if someone will want more info or logs.

It works in the evening and when I came to work this morning it was not working. The last think I changed was "Operating system" from "Other" to "Windows 2008 R2 x64" and "Timezone" to "GMT" on that VM.

Does the VM have IP details defined? If this situation occurs, can you try to ping from inside the VM and see (with tcpdump) it the traffic flows on the bridge, and if it gets from the bridge to the actual interface (or bond)? Also, there might be something interesting from the vdsm log about network changes?

----- Original Message -----

Dne 28.11.2013 13:11, Mike Kolesnik napsal(a):

...

----- Original Message -----

...

Dne 28.11.2013 10:12, Mike Kolesnik napsal(a):

...

----- Original Message -----

...

Hello, Hi Jakub,

...

I found strange problem in vdsm 4.13.0 with oVirt 3.3.1. After 30 hours from update (ovirt -> 3.3.1, vdsm -> 4.13) my one vlan network does not work. There is nothing in log, I can see network trafic from not working vlan on node in tcpdump, but it not come to VM. Did it stop working all of the sudden or did you change something?

...

After vdsm restart on node it works like expected. Seems like VDSM restored configuration to last known config on that host, maybe there's something in vdsm.log to indicate what went wrong.

...

I take a look in to oVirt -> Host -> Network Interface and all nodes has bond interface changed from "Mode 4 Dynamic link agregation" to "Custom: mode=4" and Custom mode is filled by "mode=4" Mabe it is the same, but who change that? This is a know behavior but it's still mode 4 it's just the way it displays it.

The default bond modes changed from previous versions to be with miimon=100 so if you add that to your bond definition (or just select mode 4 again) then you should see the default mode 4 selected in the list.

...

...

...

...

I can have one node in this status if someone will want more info or logs.

It works in the evening and when I came to work this morning it was not working. The last think I changed was "Operating system" from "Other" to "Windows 2008 R2 x64" and "Timezone" to "GMT" on that VM.

Does the VM have IP details defined? If this situation occurs, can you try to ping from inside the VM and see (with tcpdump) it the traffic flows on the bridge, and if it gets from the bridge to the actual interface (or bond)?

Also, there might be something interesting from the vdsm log about network changes? Hi,

we found problem. It was in network infrastructure. So nothink with oVirt. I am sorry ;-)

Great, I'm happy you found the problem, don't be sorry :)