Nodes not starting with strange Open vSwitch in logs:
Posted: Sun Sep 22, 2019 4:50 pm
Hello community.
I'm trying to run eve-ng 2.0.3-95 (Community. version) in a rack server running Oracle linux 3.10.0-862.11.6 but I'm getting an OVS error when trying to start any node via eve-ng gui. GUI says node started but it stays down and greyed out.
This happens for vIOL, vCSR, vQFX.. so i guess is not node related. i'm running a bit out of ideas as looks to be something to do with openvswitch, maybe? With my shallow knowledge, appears like the ovs cannot find the vunl node when trying to create the bridge.
Doing 'ip link' in the eve-ng (ubuntu) cli, indeed I see plenty of vunl interfaces but all down
Sep 22 19:06:43 eve-ng ovs-vsctl: ovs|00002|db_ctl_base|ERR|no port named vunl0_2_16
Sep 22 19:06:43 eve-ng ovs-vsctl: ovs|00002|db_ctl_base|ERR|no port named vunl0_2_32
Sep 22 19:06:43 eve-ng ovs-vsctl: ovs|00002|db_ctl_base|ERR|no port named vunl0_2_48
This same ova, in my Dell laptop, runs just fine (fedora linux 29)I have tested this with same ova in KVM and VirtualBox hypervisors, same issue.
These are the specs from the host:
Processor Architecture: x86 64-bit
Processors: 2 / 2 (Installed / Maximum)
Processor Summary: Two Intel Xeon Processor Scalable Family
Installed RAM Size: 512 GB
Guest (these are the current settings, but problem is the same with much higher cpu and memory allocated)
# vboxmanage showvminfo "EVE Community VM" | egrep -i "memory|cpu"
Memory size 8192MB
CPU exec cap: 100%
CPUProfile: host
Number of CPUs: 4
CPUID Portability Level: 0
CPUID overrides: None
Configured memory balloon size: 0MB
Thanks a lot in advance for your help!
I'm trying to run eve-ng 2.0.3-95 (Community. version) in a rack server running Oracle linux 3.10.0-862.11.6 but I'm getting an OVS error when trying to start any node via eve-ng gui. GUI says node started but it stays down and greyed out.
This happens for vIOL, vCSR, vQFX.. so i guess is not node related. i'm running a bit out of ideas as looks to be something to do with openvswitch, maybe? With my shallow knowledge, appears like the ovs cannot find the vunl node when trying to create the bridge.
Doing 'ip link' in the eve-ng (ubuntu) cli, indeed I see plenty of vunl interfaces but all down
Sep 22 19:06:43 eve-ng ovs-vsctl: ovs|00002|db_ctl_base|ERR|no port named vunl0_2_16
Sep 22 19:06:43 eve-ng ovs-vsctl: ovs|00002|db_ctl_base|ERR|no port named vunl0_2_32
Sep 22 19:06:43 eve-ng ovs-vsctl: ovs|00002|db_ctl_base|ERR|no port named vunl0_2_48
This same ova, in my Dell laptop, runs just fine (fedora linux 29)I have tested this with same ova in KVM and VirtualBox hypervisors, same issue.
These are the specs from the host:
Processor Architecture: x86 64-bit
Processors: 2 / 2 (Installed / Maximum)
Processor Summary: Two Intel Xeon Processor Scalable Family
Installed RAM Size: 512 GB
Guest (these are the current settings, but problem is the same with much higher cpu and memory allocated)
# vboxmanage showvminfo "EVE Community VM" | egrep -i "memory|cpu"
Memory size 8192MB
CPU exec cap: 100%
CPUProfile: host
Number of CPUs: 4
CPUID Portability Level: 0
CPUID overrides: None
Configured memory balloon size: 0MB
Thanks a lot in advance for your help!