Hello!
I have a problem with two VM´s on a HP Proliant ML350p Gen8, that has Windows Server 2012R2 operating system with Hyper-V role installed.
Both VM´s are also 2012R2, one is DC/DNS/FILE/PRINT and the other one is Exchange 2013.
The problem itself is that, when Windows backup (wbengine.exe - Block level backup engine service) inside VM starts, it uses almost maximum CPU resource, so the VM´s CPUs are 100% used. Backup itself is not working, there is no network trafic. Only way to get rid of 100% CPU usage is to restart the server.
When i start backup manually (Backup once - scheduled backup options), it is working and the CPU usage is around 20-30%.
One VM (DC/DNS/FILE/PRINT) has 2x vCPU and Exchange has 4x vCPU.
The problem started suddenly, the backup were performing slowly and failed.
Backups are scheduled on different time. Host level Hyper-V backup of VM´s works without any problem.
The backup location is Synology DS414 NAS and the servers are set up using iSCSI, for connecting separate backup lun´s for each server.
The Hyper-V host and both VM´s are patched up-to-date. Ofcourse i have tried a lot of restarts, looking event log, VSSADMIN - cannot find any errors...
The network setup:
Hyper-V host has a team of two physical adapters, that is set up with a external virtual switch (shared with management operating system) and used by the VM`s normal network trafic.
Hyper-V host itself is using separte physcial network card, for connecting with iSCSI to Synology NAS.
I tried to set up another physical network card on the host and create a new virtual swich, add extra virtual nic to VM`s and use this for iSCSI traffic for Windows backup and it didnt make things better.
I need to find a solution, because i cannot restart the server every night for the backups to work.