[Openstack] Services DOWN - why?

Turbo Fredriksson turbo at bayour.com
Fri Feb 3 16:07:02 UTC 2017


> On 28 Dec 2016, at 20:22, turbo at bayour.com wrote:
> 
>>> What in Openstack is it that records if a service is UP or DOWN?
>>>  
>>> Every time I have to turn of my whole stack, it takes lots and lots of
>>> trials, restarts of services etc etc to get things up and running.
>>>  
>>> Looking at the logs, there’s nothing wrong with the service(s), but
>>> yet they’re recorded as DOWN.
> 
> On 29 Dec 2016, at 08:22, chagg at foxmail.com wrote:
> 
>> servicegroup.drivers.db.DbDriver#is_up
>> Nova uses update_at culomn to decide if the service is up.

It’s been a while, but now I’m there again!

Dumbasses in the shop downstairs was playing with the power
lines, so all power in my apartment died as well :(.


I’ve taken up my two primaries (Nova, Neutron, Cinder etc masters)
but only one “comes up”. The second primary refuses to change to
“UP”.

What database and table is that column located in? And why won’t
the second primarys services come up?!


I’ve checked the RabbitMQ and MySQL cluster - all ok as far as I can
tell.

I’ve restarted all Nova, Neutron and Cinder services on the secondary
master but can’t see anything. The logs looks all ok but the services are
still listed as “DOWN”.



More information about the Openstack mailing list