Windows Server Crashes Adding more than 1 CPU

Before posting something, READ the changelog, WATCH the videos, howto and provide following:
Your install is: Bare metal, ESXi, what CPU model, RAM, HD, what EVE version you have, output of the uname -a and any other info that might help us faster.

Moderator: mike

lalovrod
Posts: 8
Joined: Sun Aug 18, 2019 8:22 pm

Windows Server Crashes Adding more than 1 CPU

Post by lalovrod » Wed Sep 11, 2019 7:06 am

Hi

I think I have faced a software defect on EVE, when I add a W2016 server, it only boots correctly when I select only 1 CPU, if I add more than 1 CPU simply the OS does not boot and it breaks the SSH connection that I have to the eve-ng

I think a need to tweak a QEMU option but I do not know which

Any help to use more than 1 CPU in Windows would be appreciated

EVE-NG is installed in google cloud, is the professional version.

The logs that I could connect tail -f /opt/unetlab/data/Logs/unl_wrapper.txt

A Success W2016 boot

root@eve-instance-01:/opt/unetlab/data/Logs# tail -f unl_wrapper.txt
Sep 11 05:55:11 Sep 11 05:55:11 Online Check state: Valid
Sep 11 05:55:11 Sep 11 05:55:11 LOG: will start node 39 0 300 1 57667
Sep 11 05:55:11 INFO: tunctl -u 33831 -g root -t vunl1_0_39_0 2>&1
Sep 11 05:55:11 INFO: sysctl -w net.ipv6.conf.vunl1_0_39_0.disable_ipv6=1
Sep 11 05:55:11 INFO: vunl1_0_39_0 -> 0,0,0,0
Sep 11 05:55:11 ERROR: QEMU Arch is not set (80015).
Sep 11 05:55:11 INFO: CWD is /opt/unetlab/tmp/0/672087a3-4ee3-4b00-b4eb-14c9e337676b/39
Sep 11 05:55:11 INFO: starting /opt/unetlab/wrappers/qemu_wrapper -C 57667 -T 0 -D 39 -t "Winserver39" -F /opt/qemu-2.2.0/bin/qemu-system-x86_64 -d 0 -x -- -device e1000,netdev=net0,mac=50:00:00:27:00:00 -netdev tap,id=net0,ifname=vunl1_0_39_0,script=no -smp 1,sockets=1,cores=1,threads=1 -m 8192 -name Winserver39 -uuid cf33c82e-bd77-4da2-9795-f8746bf4787f -qmp unix:./qmp-sock,server,nowait -monitor unix:./mon-sock,server,nowait -vnc :51767 -hda hda.qcow2 -machine type=pc-1.0,accel=kvm -cpu qemu64,+fsgsbase -vga std -usbdevice tablet -boot order=dc -drive file=/opt/qemu/share/qemu/virtio-win-drivers.img,index=1,if=floppy,readonly > /opt/unetlab/tmp/0/672087a3-4ee3-4b00-b4eb-14c9e337676b/39/wrapper.txt 2>&1 & echo $!
Sep 11 05:55:11 INFO: started process is7858


Sep 11 05:57:53 Sep 11 05:57:53 Online Check state: Valid
Sep 11 05:57:53 Sep 11 05:57:53 DEBUG int - net: vunl1_0_39_0 - pnet1
Sep 11 05:57:53 LOG: exec : brctl addif pnet1 vunl1_0_39_0 2>&1
Sep 11 05:57:53 Sep 11 05:57:53 Online Check state: Valid
Sep 11 05:57:53 INFO: vunl1_0_39_0 -> 0,0,0,0


A nonsuccess boot

CRASH
Sep 11 06:07:35 Sep 11 06:07:35 Online Check state: Valid
Sep 11 06:07:35 Sep 11 06:07:35 LOG: will start node 39 0 300 1 41475
Sep 11 06:07:35 INFO: tunctl -u 33831 -g root -t vunl1_0_39_0 2>&1
Sep 11 06:07:35 INFO: sysctl -w net.ipv6.conf.vunl1_0_39_0.disable_ipv6=1
Sep 11 06:07:35 INFO: vunl1_0_39_0 -> 0,0,0,0
Sep 11 06:07:35 LOG: exec : brctl addif pnet1 vunl1_0_39_0 2>&1
Sep 11 06:07:35 ERROR: QEMU Arch is not set (80015).
Sep 11 06:07:35 INFO: CWD is /opt/unetlab/tmp/0/672087a3-4ee3-4b00-b4eb-14c9e337676b/39
Sep 11 06:07:35 INFO: starting /opt/unetlab/wrappers/qemu_wrapper -C 41475 -T 0 -D 39 -t "Winserver39" -F /opt/qemu-2.2.0/bin/qemu-system-x86_64 -d 0 -x -- -nographic -device e1000,netdev=net0,mac=50:00:00:27:00:00 -netdev tap,id=net0,ifname=vunl1_0_39_0,script=no -device e1000,netdev=net1,mac=50:00:00:27:00:01 -netdev user,id=net1,hostfwd=tcp::41475-:3389,net=169.254.1.100/30,dhcpstart=169.254.1.100,restrict=on -smp 4,sockets=1,cores=4,threads=1 -m 8192 -name Winserver39 -uuid cf33c82e-bd77-4da2-9795-f8746bf4787f -qmp unix:./qmp-sock,server,nowait -monitor unix:./mon-sock,server,nowait -hda hda.qcow2 -machine type=pc-1.0,accel=kvm -cpu qemu64,+fsgsbase -vga std -usbdevice tablet -boot order=dc -drive file=/opt/qemu/share/qemu/virtio-win-drivers.img,index=1,if=floppy,readonly > /opt/unetlab/tmp/0/672087a3-4ee3-4b00-b4eb-14c9e337676b/39/wrapper.txt 2>&1 & echo $!
Sep 11 06:07:35 INFO: started process is18929
Sep 11 06:07:35 INFO: vunl1_0_39_0 -> 0,0,0,0
packet_write_wait: Connection to 34.69.174.44 port 22: Broken pipe

Uldis (UD)
Posts: 5067
Joined: Wed Mar 15, 2017 4:44 pm
Location: London
Contact:

Re: Windows Server Crashes Adding more than 1 CPU

Post by Uldis (UD) » Wed Sep 11, 2019 8:31 am

No need more cpu for it, RAM raising will do better performance
Just did Server 2019 test with x4 CPU and 16G RAM
all nice
You do not have the required permissions to view the files attached to this post.

lalovrod
Posts: 8
Joined: Sun Aug 18, 2019 8:22 pm

Re: Windows Server Crashes Adding more than 1 CPU

Post by lalovrod » Wed Sep 11, 2019 5:28 pm

Hi, I have taken some screenshots about the Settings of the VM

First Success Using QEMU Version 2.2.0 , QEMU Arch x86_64 and QEMU Nic e1000, W2016 Boots without any problem with only 1 CPU
SUCCESS.png
However, using the same settings changing only the number of CPUs it crashes
FAIL.png
I changed the QEMU Version to tpl 2.0.2, QEMU ARCH to tpl x86_64 and the NIC to tpl e1000 as many videos on the Internet shows that Windows boots without any problem, using only 1 CPU it boots correctly
Success2.png
If I change the number of CPUS, with the same settings it does not boot.
You do not have the required permissions to view the files attached to this post.

Uldis (UD)
Posts: 5067
Joined: Wed Mar 15, 2017 4:44 pm
Location: London
Contact:

Re: Windows Server Crashes Adding more than 1 CPU

Post by Uldis (UD) » Wed Sep 11, 2019 5:59 pm

Qemu indeed keep older
here is my 2019
You do not have the required permissions to view the files attached to this post.

lalovrod
Posts: 8
Joined: Sun Aug 18, 2019 8:22 pm

Re: Windows Server Crashes Adding more than 1 CPU

Post by lalovrod » Wed Sep 11, 2019 10:08 pm

The same scenario happen when I added a W10 Client.....

Uldis (UD)
Posts: 5067
Joined: Wed Mar 15, 2017 4:44 pm
Location: London
Contact:

Re: Windows Server Crashes Adding more than 1 CPU

Post by Uldis (UD) » Thu Sep 12, 2019 7:12 am

Come to EVE live chat to talk:
https://www.eve-ng.net/live-helpdesk
Use your gmail account or create new to join in chat.
my chatname
EVE-NG-UD

Probably it is your Win image issue, wrongly made

Uldis

Win10 belo
x2CPU and 8G RAM
You do not have the required permissions to view the files attached to this post.

lalovrod
Posts: 8
Joined: Sun Aug 18, 2019 8:22 pm

Re: Windows Server Crashes Adding more than 1 CPU

Post by lalovrod » Thu Sep 12, 2019 1:17 pm

From unl_wrappers log files, QEMU Version used is 2.0.2

Sep 12 12:44:09 ERROR: QEMU Arch is not set (80015).
Sep 12 12:44:10 INFO: CWD is /opt/unetlab/tmp/0/672087a3-4ee3-4b00-b4eb-14c9e337676b/39
Sep 12 12:44:10 INFO: starting /opt/unetlab/wrappers/qemu_wrapper -C 59333 -T 0 -D 39 -t "W2016" -F /opt/qemu-2.0.2/bin/
qemu-system-x86_64 -d 0 -x -- -nographic -device e1000,netdev=net0,mac=50:00:00:27:00:00 -netdev tap,id=net0,ifname=vun
l1_0_39_0,script=no -device e1000,netdev=net1,mac=50:00:00:27:00:01 -netdev user,id=net1,hostfwd=tcp::59333-:3389,net=16
9.254.1.100/30,dhcpstart=169.254.1.100,restrict=on -smp 1,sockets=1,cores=1,threads=1 -m 4096 -name W2016 -uuid cf33c82e
-bd77-4da2-9795-f8746bf4787f -qmp unix:./qmp-sock,server,nowait -monitor unix:./mon-sock,server,nowait -hda hda.qcow2
-machine type=pc-1.0,accel=kvm -cpu qemu64,+fsgsbase -vga std -usbdevice tablet -boot order=dc -drive file=/opt/qemu/sha
re/qemu/virtio-win-drivers.img,index=1,if=floppy,readonly > /opt/unetlab/tmp/0/672087a3-4ee3-4b00-b4eb-14c9e337676b/39/w
rapper.txt 2>&1 & echo $!
Sep 12 12:44:10 INFO: started process is6768

The CPUS assigned are not from 3.33 GHZ (thank all for the screenshots, but 2.8 GHZ
Screen Shot 2019-09-12 at 8.03.57 AM.png
Screen Shot 2019-09-12 at 8.03.07 AM.png
You do not have the required permissions to view the files attached to this post.

lalovrod
Posts: 8
Joined: Sun Aug 18, 2019 8:22 pm

Re: Windows Server Crashes Adding more than 1 CPU

Post by lalovrod » Thu Sep 12, 2019 1:29 pm

Its frustrating I tried to use QEMU Version 3 on the Settings and it crashes once again.

The problem is that when I boot and Windows crashes. all the lab nodes that are running also crash. They simply change from blue to gray and I could not access any of them

Please any help would be appreciated

Uldis (UD)
Posts: 5067
Joined: Wed Mar 15, 2017 4:44 pm
Location: London
Contact:

Re: Windows Server Crashes Adding more than 1 CPU

Post by Uldis (UD) » Thu Sep 12, 2019 5:07 pm

go to eve live chat,
link i sent you earlier...
to be honest it looks like you simply screwed all your lab with that win nodes test, in result crash all.
another reason could be simply weak cpu and ram to keep heavy lab
Uldis

lalovrod
Posts: 8
Joined: Sun Aug 18, 2019 8:22 pm

Re: Windows Server Crashes Adding more than 1 CPU

Post by lalovrod » Tue Sep 17, 2019 12:52 am

I will join tomorrow to the live chat, today I capture this log file, when the Windows crashes

Sep 15 19:43:47 eve-instance-01 systemd[1]: Started OS configuration for EVE-NG.
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Running Virtual Environment...
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Check new disks
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: New disk sda found
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Device /dev/sda not found (or ignored by filtering).
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Detecting VG....
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Please enter a physical volume path.
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Run `vgextend --help' for more information.
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Detecting ROOT FS.../dev/sda1
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Path required for Logical Volume "sda1".
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Please provide a volume group name
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Run `lvextend --help' for more information.
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: Resizing ROOT FS
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: resize2fs 1.42.13 (17-May-2015)
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: The filesystem is already 131071739 (4k) blocks long. Nothing to do!
Sep 15 19:43:47 eve-instance-01 systemd-udevd[5322]: error opening ATTR{/sys/devices/pci0000:00/0000:00:03.0/virtio0/host0/target0:0:1/0:0:1:0/block/sda/sda1/queue/scheduler} for writing: No such file or directory
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: net.bridge.bridge-nf-call-arptables = 0
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: net.bridge.bridge-nf-call-ip6tables = 0
Sep 15 19:43:47 eve-instance-01 ovfstartup.sh[5249]: net.bridge.bridge-nf-call-iptables = 0

Post Reply