Windows VM not install High CPU

[wimp9849]

Hi, new intallation. HP ml10 gen9 2x1Tb sata zfs mirror.

When i create a new windows vm following the video tutorial the install iso never stop to load, i can't reach the installation wizard. Tried windows 2016 and 10. Cpu is costant at 25% (4vcore assignet do vm, so seem it is using only 1 core). waited 30 min but the there is always the black screen with the dots turning in circle endless and the cpu is always costant at 25%.

Tried using virtio scsi, ide, sata. Tried also to load only the iso without a disk attached. Balooning on and off. Always the same

Tried a centos 7 container and a centos 7 vm and is very fast, installed a KVM Centos 7 machine in 8 mins and it took less than a min from boot to installation wizard screen, so the problem is only with windows.

What could it be?

 

[tom]

Faulty ISO?

 

[wimp9849]

Faulty ISO?

Tried 2 different ISO. Windows 2016 and Windows 10

 

[tom]

Tried 2 different ISO. Windows 2016 and Windows 10

you should test if the ISO are correctly available, e.g. verify the sha256sum or use similar tool.

1. and post your VM config:
> qm config VMID (of your windows VM)

2. and the output of:
> pveversion -v

 

[wimp9849]

you should test if the ISO are correctly available, e.g. verify the sha256sum or use similar tool.

1. and post your VM config:
> qm config VMID (of your windows VM)



2. and the output of:
> pveversion -v

1

root@pve01:~# qm config 102
agent: 1
balloon: 0
boot: cdn
bootdisk: scsi1
cores: 4
ide2: local:iso/SW_DVD9_NTRL_Windows_Svrs_2016_Italian_2_Std_DC_FPP_OEM_X21-22571.ISO,media=cdrom,size=5555902K
memory: 4096
name: Win2016
net0: virtio=2E:56:A0:9A:1C:AA,bridge=vmbr0
numa: 0
ostype: win10
scsi0: local-zfs:vm-102-disk-1,discard=on,size=500G
scsihw: virtio-scsi-pci
smbios1: uuid=2c953001-7cdf-4587-9492-60ad7337fb97
sockets: 1

2

proxmox-ve: 5.1-32 (running kernel: 4.13.13-2-pve)
pve-manager: 5.1-41 (running version: 5.1-41/0b958203)
pve-kernel-4.13.13-2-pve: 4.13.13-32
libpve-http-server-perl: 2.0-8
lvm2: 2.02.168-pve6
corosync: 2.4.2-pve3
libqb0: 1.0.1-1
pve-cluster: 5.0-19
qemu-server: 5.0-18
pve-firmware: 2.0-3
libpve-common-perl: 5.0-25
libpve-guest-common-perl: 2.0-14
libpve-access-control: 5.0-7
libpve-storage-perl: 5.0-17
pve-libspice-server1: 0.12.8-3
vncterm: 1.5-3
pve-docs: 5.1-12
pve-qemu-kvm: 2.9.1-5
pve-container: 2.0-18
pve-firewall: 3.0-5
pve-ha-manager: 2.0-4
ksm-control-daemon: 1.2-2
glusterfs-client: 3.8.8-1
lxc-pve: 2.1.1-2
lxcfs: 2.0.8-1
criu: 2.11.1-1~bpo90
novnc-pve: 0.6-4
smartmontools: 6.5+svn4324-1
zfsutils-linux: 0.7.3-pve1~bpo9

 

[tom]

looks ok. did you test the ISO already?

 

[wimp9849]

i've tried disabling Sync on zfs pool and now in 10 min the install wizard show up and windows installed in less than 10 minute...

more than 30 minutes for loading the cd rom is insane even with zfs sync enabled. what could be wrong?

 

[wimp9849]

Little Update:

setting min arc at 1GB seem fixed the problem now. without setting min arc on rebooting the vms took about 15 min to restart.

options zfs zfs_arc_min=1073741824
options zfs zfs_arc_max=4294967296

 

[wimp9849]

update: Shutting down the vm and next start it tooks ages to booting like 20/25 min. Is this an expected behavior?

 

[GadgetPig]

I would double check the ML110 BIOS raid/SATA settings. I would also try a new SATA cable, and maybe a different SATA port. To better troubleshoot (and if this is just a test server) start Proxmox install on the ML110 with a single SATA/SAS disk and regular ext4. When you create a Windows VM,start with just 1 or 2 vcpu. 20/25 min booting (assuming it's the only VM running) is not normal.

 

[wimp9849]

I would double check the ML110 BIOS raid/SATA settings. I would also try a new SATA cable, and maybe a different SATA port. To better troubleshoot (and if this is just a test server) start Proxmox install on the ML110 with a single SATA/SAS disk and regular ext4. When you create a Windows VM,start with just 1 or 2 vcpu. 20/25 min booting (assuming it's the only VM running) is not normal.

The server is 1 months old, and before yesterday was running hyper v with 2 vms without problem using internal "hw" raid. Is an HP ML 10 not ML110.
In bios i set sata to AHCI, enabled CSM in legacy only mode ( no uefi), am i missing some setting? I can try without ZFS in any case, probably 8 GB ram is not enough for testing and probably the problem is here and not in only using 2 drive.

Thanks for your support.

 

[GadgetPig]

Yes 8GB is probably the issue. ZFS requires lots of memory to be optimal. So if you have 8GB memory, about 50%/4GB is usually reserved for ZFS. I recommend trying again with regular ext4, or adding an extra 8GB ram.

 

[GadgetPig]

@tuonazzurro here is a similar post about ZFS memory requirements:

https://forum.proxmox.com/threads/proxmox-ve-new-server-high-io-delay.39162/#post-193737

 

[wimp9849]

Yes 8GB is probably the issue. ZFS requires lots of memory to be optimal. So if you have 8GB memory, about 50%/4GB is usually reserved for ZFS. I recommend trying again with regular ext4, or adding an extra 8GB ram.

Thanks for your answers, i'll give it a try with 16 GB of Ram and setting ARC max to 5 GB (4gb +1GB/Tb)