[Openstack-operators] metadata-api 500 errors
aleonhardt.py at gmail.com
Thu Jan 15 17:17:48 UTC 2015
Hmmm, I did do some playing with security groups, but even the last network
(network5) that I created was unchanged from "default" - maybe it has
something to do with that, I'm testing a brand new tenant/project now with
a similar setup, except I'm setting up all the networking / networks first
On Thu Jan 15 2015 at 17:12:47 Edgar Magana <edgar.magana at workday.com>
> Did you change the security groups?
> I will try to compare the iptables configuration for each name space
> From: Alex Leonhardt <aleonhardt.py at gmail.com>
> Date: Thursday, January 15, 2015 at 9:03 AM
> To: Edgar Magana <edgar.magana at workday.com>, openstack-operators <
> openstack-operators at lists.openstack.org>
> Subject: Re: [Openstack-operators] metadata-api 500 errors
> Hi Edgar,
> that's the crazy thing - so all the gre tunnels are up, I can see them in
> openvswitch and also can see that there are some openflow rules applied.
> I've craeted VMs on every hypervisor (including the controller, as it's a
> test install) on network1 (192.168.1.0), every VM (and that is still the
> case now) started there works just fine and gets the metadata as expected,
> the same for network2 (192.168.2.0).
> the issue only appeared after I created network3 (192.168.3.0), VMs
> there (tried again all 3 hypervisors) get a 500 errror instead of the
> expected metadata files/json. The same for any / all other networks I
> created after (network4 and 5).
> On the VM all I can see is this:
> 2015-01-15 17:02:57,310 - url_helper.py[WARNING]: Calling 'http://169.254.169.254/2009-04-04/meta-data/instance-id' failed [0/120s]: bad status code 
> 2015-01-15 17:02:58,509 - url_helper.py[WARNING]: Calling 'http://169.254.169.254/2009-04-04/meta-data/instance-id' failed [1/120s]: bad status code 
> On Thu Jan 15 2015 at 16:53:31 Edgar Magana <edgar.magana at workday.com>
>> Did you follow the networking recommendations:
>> It will ell you if you write your own topology and complete a packet
>> trace to find out the issue.
>> Make sure all tunnels are established between your three nodes.
>> From: Alex Leonhardt <aleonhardt.py at gmail.com>
>> Date: Thursday, January 15, 2015 at 7:45 AM
>> To: openstack-operators <openstack-operators at lists.openstack.org>
>> Subject: [Openstack-operators] metadata-api 500 errors
>> i've got a test openstack install with 3 nodes, using gre tunneling --
>> initially it all worked fine, but, after creating > 2 networks, VMs in
>> networks 3,4,5 do not seem to get the metadata due to it erroring with 500
>> errors. whilst this is happening, VMs in networks 1 and 2 are still working
>> fine and can be provisioned OK.
>> anyone seen something similar or ideas on how to go about
>> troubleshooting this ? I got a tcpdump from the VM but as it does get to
>> the metadata api, am not sure where the issue is (especially since other
>> VMs in other Networks work just fine)
>> any ideas ?
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the OpenStack-operators