[Starlingx-discuss] Compute boot-loop

Michel Thebeau michel.thebeau at windriver.com
Mon Nov 5 14:55:22 UTC 2018


Hi Volker,

The puppet log is pointing at vswitch.  I'll ask one of the networking 
guys to look at this thread.

2018-11-05T11:16:26.170 Notice: 2018-11-05 11:16:26 +0000 
/Stage[main]/Platform::Vswitch::Ovs/Platform::Vswitch::Ovs::Device[0000:01:00.0]/Exec[ovs-bind-device: 
0000:01:00.0]/returns: Error: bind failed for 0000:01:00.0 - Cannot bind 
to driver vfio-pci
2018-11-05T11:16:26.173 Notice: 2018-11-05 11:16:26 +0000 
/Stage[main]/Platform::Vswitch::Ovs/Platform::Vswitch::Ovs::Device[0000:01:00.0]/Exec[ovs-bind-device: 
0000:01:00.0]/returns: Error: unbind failed for 0000:01:00.0 - Cannot 
open /sys/bus/pci/drivers//unbind
2018-11-05T11:16:26.175 Error: 2018-11-05 11:16:26 +0000 
dpdk-devbind.py --bind=vfio-pci 0000:01:00.0 returned 1 instead of one 
of [0]

===

PS.  It is normal for ifconfig and ip commands to not show the 
interfaces configured for platform data.

M


On 2018-11-05 8:56 a.m., von Hoesslin, Volker wrote:
> build instructions: https://docs.starlingx.io/developer_guide/ (2018-11-02)
> install instruction: https://docs.starlingx.io/installation_guide/dedicated_storage.html#dedicated-storage
>
> the installation runs without problems, only at the very end the unlocking of the compute-nodes ends in said endless loop. here are the last ~500kb of puppet.log from the compute-node:
>
> https://pastebin.com/wsSUY741
>
> it seems that there are problems with the network interface eno1. the output of "ip a" only shows eno2, eno1 is not started.
> did i understand correctly that the "data" interface at the compute-nodes is the interface via which the "provider-networks" can be reached (via VLAN) ? i refer to this line from the doku:
>
> system host-if-modify -p providernet-a -c data compute-0 eno1
>
> here is an overview of my network, basically very primitive and i think also in the sense of starlingX/openstack quite classic.
>
> https://docdro.id/0CRp2OA
>
> i am grateful for any help!
>
> Volker...
>
> -----Ursprüngliche Nachricht-----
> Von: Michel Thebeau [mailto:michel.thebeau at windriver.com]
> Gesendet: Donnerstag, 1. November 2018 17:29
> An: von Hoesslin, Volker; 'starlingx-discuss at lists.starlingx.io'
> Betreff: Re: [Starlingx-discuss] Compute boot-loop
>
> Hi Volker,
>
> Could you describe the instructions you were using to setup startlingx
> (were you following a wiki online)?  Did you recheck the instructions
> versus what is actually configured on the computes?
>
> How many cpus do your computes have?
>
> Which build are you using?  Did you build your own?
>
> When reboot-loop happens on a node often the error will be presented on
> the console during boot or could be inferred from console, sometimes
> referring to puppet logs of the node /var/log/puppet/
>
> M
>
>
> On 2018-11-01 6:45 a.m., von Hoesslin, Volker wrote:
>> Hello, there,
>> i have now built a complete stack and believe i did everything right. after unlocking the compute nodes (two pieces), the nodes come into an endless boot loop.
>> Error messages are not really noticeable, except:
>> compute-X 'ceilometer-polling' process has failed. Auto recovery in progress
>>
>> any suggestions? Greez & thx, volker...
>>
>>
>> _______________________________________________
>> Starlingx-discuss mailing list
>> Starlingx-discuss at lists.starlingx.io
>> http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss
>
> _______________________________________________
> Starlingx-discuss mailing list
> Starlingx-discuss at lists.starlingx.io
> http://lists.starlingx.io/cgi-bin/mailman/listinfo/starlingx-discuss




More information about the Starlingx-discuss mailing list