hi, after an power outage my server tells me similar problems... I am curious what kind of answers will come ;)


volker...


Von: Pratik M. <pvmpublic@gmail.com>
Gesendet: Freitag, 14. August 2020 12:31:08
An: starlingx-discuss@lists.starlingx.io
Betreff: [URL wurde verändert] [Starlingx-discuss] Debugging operational state disabled
 
Externe E-Mail! Öffnen Sie nur Links oder Anhänge von vertrauenswürdigen Absendern! 
Hi
I have a R3.0 AIO-DX cluster. I have been able to get into a state twice where I try some cluster-wide config change, do a lock and unlock of controller-1 and it ends up being operational=disabled. A force lock-unlock or reboot does not help.

I have hit this twice:
1. Moving OAM_IF to a bonded interface
2. Trying to add OVS-DPDK. I did this (appx):

system modify --vswitch_type ovs-dpdk
system host-lock controller-1
system host-cpu-modify -f vswitch -p0 1 controller-1
system host-memory-modify -f vswitch -1G 1 controller-1 0
system host-memory-modify -f vswitch -1G 1 controller-1 1
system host-memory-modify controller-1 0 -1G 10
system host-memory-modify controller-1 1 -1G 10
system host-unlock controller-1

I was then planning to do swact from controller-0 and repeat the steps. But I end up with controller-1 as disabled.

I have a couple of pods in CrashLoopBackOff, and error logs w/ ovs-ofctl complaining about interfaces that do not exist.
https://sis-schwerin.de/externer-link/?href=https://gist.github.com/pvmpublic/2f9a7180a5aef59de09aac9877302720 
 
How do I debug this?

Thanks in advance
Pratik