On Tue, Sep 4, 2018 at 9:02 AM Edward Haas <ehaas@redhat.com> wrote:

Hello Florian,

Thanks for checking the patch and posting the bug.

You need to restart vdsmd and supervdsmd.

It should not affect running VM/s, but you always have a risk that something unexpected can happen. Perhaps try it on a host and then proceed with others.

Thanks,

Edy.

I'm having similar problem in a 3 hosts oVirt test cluster with these notifications every day on 1Gbit adapters.

I have bond0 on em1 and em2 and then bondo.65, bond0.68, bond0.167 vlans defined for the VMs

I get these warnings 

Message:Host ov300 has network interface which exceeded the defined threshold [95%] (em1: transmit rate[98%], receive rate [0%])

when actually I think the 3 VMs running on this host generate few MB/s of traffic

I applied the changes to the 3 hosts. 

I notice that due to dependencies it is sufficient to restart supervdsmd and then also vdsmd will be automatically restarted, correct?

In my case for each of the 3 hosts, after restarting supervdsmd I got messages like these, but without impacts on runnign VMs

VDSM ov300 command GetStatsAsyncVDS failed: Broken pipe 9/4/18 9:07:52 AM

Host ov300 is not responding. It will stay in Connecting state for a grace period of 61 seconds and after that an attempt to fence the host will be issued. 9/4/18 9:07:52 AM

No faulty multipath paths on host ov300 9/4/18 9:07:58 AM

Executing power management status on Host ov300 using Proxy Host ov200 and Fence Agent ipmilan:10.10.193.103. 9/4/18 9:07:58 AM

Status of host ov300 was set to Up. 9/4/18 9:07:58 AM

Host ov300 power management was verified successfully. 9/4/18 9:07:58 AM

Please note that when doing on SPM host you could also get these:

VDSM ov301 command SpmStatusVDS failed: Broken pipe 9/4/18 9:10:00 AM

Host ov301 is not responding. It will stay in Connecting state for a grace period of 81 seconds and after that an attempt to fence the host will be issued. 9/4/18 9:10:00 AM

Invalid status on Data Center MYDC. Setting Data Center status to Non Responsive (On host ov301, Error: Network error during communication with the Host.). 9/4/18 9:10:00 AM

with reassignment of SPM role:

VDSM command GetStoragePoolInfoVDS failed: Heartbeat exceeded 9/4/18 9:10:12 AM

Storage Pool Manager runs on Host ov200 (Address: ov200), Data Center MYDC. 9/4/18 9:10:14 AM

Probably safer to manually move the SPM before restarting supervdsmd on that host.

Let's see this evening if I will get any message about thresholds.

BTW: one question. I see in the code iface.Type.NIC and now also iface.Type.BOND. Don't you think that you should manage also the network teaming option available in RH EL 7, as described here:

https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/networking_guide/ch-configure_network_teaming

? 

This only if it is supported to use the new network teaming implementation in oVirt, and I'm not sure about it...

Thanks,

Gianluca