If the VMs are starting properly and they don't use cloudinit - then the issue is not oVirt specific, but guest specific (Linux/Windows depending on guest OS).

So your should check:
1. Does your host have any Netwprks out of sync (Host's network tab)
If yes - put the server into maintenance and fix the issue (host's network tab)

2. Check each VM's configuration if it is defined to use CloudInit -> if yes, verify that cloudinit's service is running on the guest

3. Verify each problematic guest's network settings. If needed, set a static IP and try to ping another IP from the same subnet/Vlan .

On Dec 25, 2019 11:41, lifuqiong@sunyainfo.com wrote:

Dear All:
My ovirt engine managed two vdsms with nfs storage on another nfs server, it worked fine about three months.
One of host (host_1.3; ip:172.18.1.3) was created about 16 vms, but the host_1.3 was shutdown unexpectly about 2019-12-23 16:11, when re-started host and vms; half of the vms was lost some configure or changed , such as lost theirs ip etc.(the vm name is 'zzh_Chain49_ACG_M' in the vdsm.log)

the vm zzh_Chain49_ACG_M was create by rest API through vm's templated . the template is 20.1.1.161; the vm zzh_Chain49_ACG_M was created by ovirt rest api and changed the ip to 20.1.1.219
by ovirt rest api. But the ip was changed to templated ip when the accident happened.

the vm's os is centos.

Hope get help from you soon, Thank you.

Mark
Sincerely.