[Openstack-operators] [kolla] Question about how Operators deploy

Steven Dake (stdake) stdake at cisco.com
Sun Feb 14 12:44:15 UTC 2016

Hey folks,

I guess we agreed to do this at the Kolla midcycle already, but the core team is mostly on vacation this week, and I was having trouble getting a discussion going on this on IRC.  The team talks faster then I can write notes :)  I really want to thank the Operators that have responded on this thread - its really fantastic to see Operators "bother" to care about a new deployment tool since past deployment tools have been so painful :)

On that note, if there are other requirements where Kolla doesn't really meet an Operational deployment needs, we would definitely like to know.  I think deployment tools are developed mostly in a vacuum and that's a real shame.  We have some operational experience on our development team, but more feedback is better then less.

Here is Sam's email to me that he asked me to forward to the operator's list.  Sam is a core reviewer in the Kolla project and doing the external ssl work.  He is having permissions problems posting to the list.

From: Sam Yaple <samuel at yaple.net<mailto:samuel at yaple.net>>
Cc: openstack-operators at lists.openstack.org<mailto:openstack-operators at lists.openstack.org>
Date: Fri, 12 Feb 2016 23:08:36 +0000
Subject: Re: [Openstack-operators] [kolla] Question about how Operators deploy

To disperse any confusion here, Kolla currently does not prevent an external VIP. It simply does not manage it. Right now you can do ssl termination and add an external vip if you as the operator set that up outside of Kolla (preferably firewalled, possibly load-balanced). What is being suggested here is that _kolla_ handle creating a user-specified public VIP and assigning the it to the keepalived/haproxy nodes and so it can properly automate ssl termination at haproxy for the external VIP.

At the mid-cycle for Liberty it was discussed to use haproxy for ssl termination, but no agreement was reached as to whether Kolla should setup a public facing VIP. It was agreed in the Mitaka (3 days ago) mid-cycle that automating an optional ssl-termination at haproxy was a desire. This means we _must_ do a second VIP as initially discussed in Liberty.

All in all what operators should get out of this is the following:

  * You can currently use external loadbalancers and firewalls and completely disable Kolla deploying and managing keepalived and haproxy for both internal and external vips
  * You can currently use Kolla to deploy keepalived and haproxy with only an internal VIP and configure an external VIP on your own firewall/loadbalancer gear
  * You will be able configure a separate VIP for your internal and a separate VIP for your external traffic with automated setup of ssl termination at haproxy (the operator must still handle firewalling the external VIP)

The first two options are available now. All three options will be at the disposal of operators very soon here. I believe Dave submitted a patch for this already (though I have not yet reviewed it).

Hopefully this clears up any confusion. Does anyone have any questions or comments about this explanation?

Sam Yaple

From: Jonathan Abdiel Gonzalez Valdebenito <jonathan.abdiel at gmail.com<mailto:jonathan.abdiel at gmail.com>>
Date: Saturday, February 13, 2016 at 6:08 PM
To: Robert Starmer <robert at kumul.us<mailto:robert at kumul.us>>, Steven Dake <stdake at cisco.com<mailto:stdake at cisco.com>>
Cc: "openstack-operators at lists.openstack.org<mailto:openstack-operators at lists.openstack.org>" <openstack-operators at lists.openstack.org<mailto:openstack-operators at lists.openstack.org>>
Subject: Re: [Openstack-operators] [kolla] Question about how Operators deploy


We also use two VIPs for tracking and security reasons a +1 for two VIP

On Sat, Feb 13, 2016, 21:00 Robert Starmer <robert at kumul.us<mailto:robert at kumul.us>> wrote:
+1 on two VIPs

On Fri, Feb 12, 2016 at 6:04 AM, Steven Dake (stdake) <stdake at cisco.com<mailto:stdake at cisco.com>> wrote:
Hi folks,

Unfortunately I won't be able to make it to the Operator midcycle because of budget constraints or I would find the answer to this question there.  The Kolla upstream is busy sorting out external ssl termination and a question arose in the Kolla community around operator requirements for publicURL vs internalURL VIP management.

At present, Kolla creates 3 Haproxy containers across 3 HA nodes with one VIP managed by keepalived.  The VIP is used for internal communication only.  Our PUBLIC_URL is set to a DNS name, and we expect the Operator to sort out how to map that DNS name to the internal VIP used by Kolla.  The way I do this in my home lab is to use NAT to NAT my public_URL from the internet (hosted by dyndns) to my internal VIP that haproxies to my 3 HA control nodes.  This is secure assuming someone doesn't bust through my NAT.

An alternative has been suggested which is to use TWO vips.  One for internal_url, one for public_url.  Then the operator would only be responsible for selecting where to to allocate the public_url endpoint's VIP.  I think this allows more flexibility without necessarily requiring NAT while still delivering a secure solution.

Not having ever run an OpenStack cloud in production, how do the Operators want it?  Our deciding factor here is what Operators want, not what is necessarily currently in the code base.  We still have time to make this work differently for Mitaka, but I need feedback/advice quickly.

The security guide seems to imply two VIPs are the way to Operate: (big diagram):

The IRC discussion is here for reference:

Thanks in Advance!

OpenStack-operators mailing list
OpenStack-operators at lists.openstack.org<mailto:OpenStack-operators at lists.openstack.org>

OpenStack-operators mailing list
OpenStack-operators at lists.openstack.org<mailto:OpenStack-operators at lists.openstack.org>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-operators/attachments/20160214/c5092de7/attachment.html>

More information about the OpenStack-operators mailing list