We've run it in a test cloud meant to identify production issues before 
supporting them on our production cloud.

We ran into a few issues that may or may not apply in your situation:

There's a security issue with the Trove Rabbitmq. The easiest way around it is 
to use the feature that lets the vm's run in a trove tenant rather then the 
users tenant. Depending on your environment, this can cause other problems with 
quota's/billing/restricted flavors though. We have not been able to deploy it 
in production due to these restrictions tied in with the security issue.

Another pain point for us was management of the guest agents. We upgraded our 
control plane for Trove, and all the existing guest agents in the vm's broke on 
us.

A third issue was incompatibility with the radosgw. Trove functionality relies 
on namespacing working right and at the time radosgw didnt fully support it. I 
think that may be fixed in Jewel, but haven't had a chance to confirm the issue 
is resolved, so there may be other issues.

Thanks,
Kevin
________________________________________
From: Masha Atakova [[email protected]]
Sent: Wednesday, March 15, 2017 12:48 AM
To: [email protected]
Subject: [Openstack-operators] Trove in production

Hi everyone,

I'm curious if anybody is using Trove in production: what's your
experience? What are the problems you've encountered and how critical
are they?

Thanks in advance for your responses


_______________________________________________
OpenStack-operators mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators

_______________________________________________
OpenStack-operators mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-operators

Reply via email to