csr1000v images do not boot after upgrade to latest EVE community 2.0.3-102

[stsfred]

hi guys,

Happy new year.
I searched for similar issues in this forum, but none of the suggestions worked for me.
I upgraded EVE to the latest version. Before the upgrade, all images worked perfectly. After upgrade, none of the CSR images boot up. When I click start, the color of the router changes to blue for a few seconds, then changes back to grey and stops.
I have 3 different CSR images installed, none of them work. XRv, IOL, IOSv images work correctly.
OS: Windows 10 Home, VMWARE workstation 15.1 player, CPU: AMD Ryzen 3900x, 64GB RAM, Eve Community edition 2.0.3-102

Code: Select all

Linux eve-ng 4.20.17-eve-ng-ukms+ #2 SMP Wed Jun 5 08:18:06 CEST 2019 x86_64 x86_64 x86_64 GNU/Linux

Code: Select all

root@eve-ng:/opt/unetlab/addons/qemu# ls -l
total 32
drwxr-xr-x 2 root root 4096 Oct 14 19:27 csr1000v-3.16.3
drwxr-xr-x 2 root root 4096 Oct 16 19:13 csr1000vng-universalk9.16.09.04.Fuji
drwxr-xr-x 2 root root 4096 Oct 14 19:19 csr1000vng-universalk9.16.6.1
drwxr-xr-x 2 root root 4096 Sep 17  2018 vios-advent-15.6.2
drwxr-xr-x 2 root root 4096 Oct 14 19:24 vios-adventerprisek9-m.SPA.157-3.T
drwxr-xr-x 2 root root 4096 Oct 14 19:30 viosl2-adventSSA-highiron
drwxr-xr-x 2 root root 4096 Oct  4 18:56 xrv9k-6.5.1
drwxr-xr-x 2 root root 4096 Oct  9  2018 xrv-k9-6.1.3
root@eve-ng:/opt/unetlab/addons/qemu# ls csr1000v-3.16.3/
virtioa.qcow2
root@eve-ng:/opt/unetlab/addons/qemu# ls -l csr1000v-3.16.3/
total 1140356
-rw-r--r-- 1 root root 1167720448 Nov  5  2017 virtioa.qcow2
root@eve-ng:/opt/unetlab/addons/qemu# ls -l csr1000vng-universalk9.16.6.1/
total 1285456
-rw-r--r-- 1 root root 1316301824 Oct  4 20:16 virtioa.qcow2
root@eve-ng:/opt/unetlab/addons/qemu# ls -l csr1000vng-universalk9.16.09.04.Fuji/
total 950280
-rw-r--r-- 1 root root 973144064 Oct 16 19:12 virtioa.qcow2

Code: Select all

root@eve-ng:/opt/unetlab/addons/qemu# kvm-ok
INFO: /dev/kvm exists
KVM acceleration can be used

I think image naming and folder naming conventions are OK.
Can someone help?
thanks.

[Uldis (UD)]

wipe node
right click/wipe

start

and afor AMD CPU use I will not guarantee anything....
Unpredictable, thats why not officially supported

EVE Community is 2.0.3.104 now
Published today

[stsfred]

Thank for your feedback Uldis.
Unfortunately, wiping didn't work either, BUT I figured out what the problem was. It may be useful for others, so:
In my case, the default Qemu parameters were the following:

eveng1.JPG

These are not correct.
After changed these parameters as shown below to this, the router booted up:

eveng2.JPG

I don't know why these settings changed, in my system, but anyway, it works now.

[Uldis (UD)]

your image is made wrongly.
CSR using new Qemu 4.1.0 now
and it works

Virtio-net-pci interface does not support dot1q subinterfaces on fuji..
The Qemu which you had selected is supperold Qemu and does not support all stuff what Fuji CSR has..

[stsfred]

according to this howto: https://www.eve-ng.net/index.php/docume ... rest-fuji/
the command line for creating the image is this:

Code: Select all

/opt/qemu-2.2.0/bin/qemu-system-x86_64  -nographic -drive file=virtioa.qcow2,if=virtio,bus=0,unit=0,cache=none -machine type=pc-1.0,accel=kvm -serial mon:stdio -nographic -nodefconfig -nodefaults -rtc base=utc -cdrom csr1000v-universalk9.16.03.03.iso -boot order=dc  -m 3072

it is qemu 2.2.0 there.

Isn't it correct? what should be the correct command for this?
changing qemu-2.2.0 to 4.1.0 doesn't work because the arguments are not valid.

Also, what about xrv9000? the command line is this:

Code: Select all

/opt/qemu-2.2.0/bin/qemu-system-x86_64  -nographic -drive file=virtioa.qcow2,if=virtio,bus=0,unit=0,cache=none -machine type=pc-1.0,accel=kvm -serial mon:stdio -nographic -nodefconfig -nodefaults -rtc base=utc -cdrom xrv9k-fullk9-x-6.3.3.iso -boot order=dc  -m 16384

It is also qemu 2.2.0. Is this correct?

[Uldis (UD)]

preparing Fuji or Denanli mage doesnt matter which qemu version you will use for installation.
Installing Fuji or Denali you can use qemu 2.2.0.
After image install is down, you simply start use to use new template new qemu and new interface vmxnet3.

Therefore will have all new features

For XRv9000 is totally same story !

And BTW the community v2.0.3-105 is OUT

[Uldis (UD)]

Just installed newest CSR Amterdam 17.01.01
Work like cram.
For install used old qemu line, but running image with new Qemu 4.1.0 and vmxnet3 NIC

[stsfred]

Thanks for the explanation.

I already upgraded to .105.

I reinstalled a 16.12.02 (Gibraltar) image exactly following all steps as show in the Howto. Still not booting up if qemu 4.1 is selected. However it boots up with qemu version 2.12 and vmxnet3 interface.
I have no idea why is this happening.



Now created a new lab and put a single CSR Gibraltar there. Corresponding logs in /opt/unatlab/data/Logs folder , php_errors.txt:

Code: Select all

[08-Jan-2020 18:44:41 Europe/Rome] Jan 08 18:44:41 DEBUG: lab saved
[08-Jan-2020 18:44:41 Europe/Rome] Jan 08 18:44:41 DEBUG: lab saved
[08-Jan-2020 18:44:52 Europe/Rome] Jan 08 18:44:52 WARNING: Attribute ignored, invalid node_uuid (40026).
[08-Jan-2020 18:44:52 Europe/Rome] Jan 08 18:44:52 DEBUG: lab saved

unl_wrapper.txt:

Code: Select all

Jan 08 18:44:59 INFO: tunctl -u unl0 -g root -t vunl0_1_0 2>&1
Jan 08 18:44:59 INFO: sysctl -w net.ipv6.conf.vunl0_1_0.disable_ipv6=1
Jan 08 18:44:59 INFO: tunctl -u unl0 -g root -t vunl0_1_1 2>&1
Jan 08 18:44:59 INFO: sysctl -w net.ipv6.conf.vunl0_1_1.disable_ipv6=1
Jan 08 18:44:59 INFO: tunctl -u unl0 -g root -t vunl0_1_2 2>&1
Jan 08 18:44:59 INFO: sysctl -w net.ipv6.conf.vunl0_1_2.disable_ipv6=1
Jan 08 18:44:59 INFO: tunctl -u unl0 -g root -t vunl0_1_3 2>&1
Jan 08 18:44:59 INFO: sysctl -w net.ipv6.conf.vunl0_1_3.disable_ipv6=1
Jan 08 18:44:59 INFO: CWD is /opt/unetlab/tmp/0/5ec8631f-b0ff-4d56-8471-8d430c3f07fb/1
Jan 08 18:44:59 INFO: starting /opt/unetlab/wrappers/qemu_wrapper -T 0 -D 1 -t "CSR11" -F /opt/qemu-4.1.0/bin/qemu-system-x86_64 -d 0 --  -nographic -device vmxnet3,netdev=net0,mac=50:00:00:01:00:00 -netdev tap,id=net0,ifname=vunl0_1_0,script=no -device vmxnet3,netdev=net1,mac=50:00:00:01:00:01 -netdev tap,id=net1,ifname=vunl0_1_1,script=no -device vmxnet3,netdev=net2,mac=50:00:00:01:00:02 -netdev tap,id=net2,ifname=vunl0_1_2,script=no -device vmxnet3,netdev=net3,mac=50:00:00:01:00:03 -netdev tap,id=net3,ifname=vunl0_1_3,script=no -smp 1 -m 4096 -name CSR11 -uuid f47bc431-6c49-4e6f-b9cf-ba8029c40f43 -drive file=virtioa.qcow2,if=virtio,bus=0,unit=0,cache=none -machine type=pc,accel=kvm -cpu host -serial mon:stdio -nographic -no-user-config -nodefaults -rtc base=utc > /opt/unetlab/tmp/0/5ec8631f-b0ff-4d56-8471-8d430c3f07fb/1/wrapper.txt 2>&1 &

Are there any other logs that may be more detailed?
The only thing I see is node uuid 40026 is invalid.

[ecze]

I suspect AMD incompatibility with qemu 4 and this image...

E.

PS: Using AMD is not officially supported. We state only that such CPU 'could' work in Native Bare metal mode...

[Uldis (UD)]

Totally no idea why you cannot use our how to and get this image ready. It is so simple.
This Time I used exactly EVE Community v 2.0.3-105, which is latest on doday, and here is result, all works like charm.

Code: Select all

cd /opt/unetlab/addons/qemu/csr1000vng-universalk9.16.12.02.Gibraltar# 

Code: Select all

/opt/qemu/bin/qemu-img create -f qcow2  virtioa.qcow2 8G

Code: Select all

/opt/qemu-2.2.0/bin/qemu-system-x86_64  -nographic -drive file=virtioa.qcow2,if=virtio,bus=0,unit=0,cache=none -machine type=pc-1.0,accel=kvm -serial mon:stdio -nographic -nodefconfig -nodefaults -rtc base=utc -cdrom csr1000v-universalk9.16.12.02.iso -boot order=dc  -m 3072

As well ecze is right, probably you are on AMD cpu and CSR simply will not run.
In our system requirements is mentioned about AMD CPU.
Below is exactly tested CSR Gibraltar version on EVE Community.