Inconsistent C&U data from OpenStack Instances


#1

Hello, I’m running Cloud Forms 4.1 using OpenStack as my cloud provider and Ceilometer as my source for chargeback data. I set this lab up about 3 months ago and everything has been generally fine with the lab, but I came back from the holidays to find that many of the provisioned instances stopped collecting data on 12/29 and none of my chargeback reports showed any data at all.

I did some troubleshooting: verified ceilometer events could be received on the appliance, verified c&u roles were properly configured, verified that the openstack configs did not change. So I did some more digging and found some instances stopped reporting data even earlier than that (CFME for one) and that a relatively new instance was reporting data. I tried refreshing most recent C&U data for several problem instances with no luck.

So my question is there a log I can consult about this? Is there a way to force C&U data to be collected? What would cause C&U to suddenly stop collecting for a particular instance?


#2

You can find the logs here: /var/www/miq/vmdb/log
Take a look at the evm.log
Do a filtered search for the following term:
ERROR

and another for: “WARN --” without quotes. (we don’t want to find WARNING)

In both searches, do a reverse order search since the log can get pretty long.


#3

@Jeff_Ekstrom, if I’m not mistaken, this was answered on a separate email list. Is that right?


#4

I think it’s figured out as a problem on the OpenStack side, but I haven’t checked Cloud Forms yet as I wanted to allow time for C&U collection.

However, I do appreciate the note above about logs as I’m not familiar with the log structure on Cloud Forms