ActiveBlog

Cloud Foundry Advisory Board Meeting - 2014 May
by Phil Whelan

Phil Whelan, May 30, 2014

CAB Meeting May 2014

Chris Ferris of IBM started off today's Cloud Foundry Community Advisory Board meeting reflecting on the recent OpenStack conference. "One of the things I observed is that it's definitely starting to trend towards discussions around containers, and Cloud Foundry and OpenStack seem to be like hand-in-glove in a lot of the conversations.".

Renat Khasanshyn from Altoros added his comments on the OpenStack Summit in the chat that "Docker conversations in the hallways heavily leaned towards Cloud Foundry" and "OpenStack ecosystem turned the corner and becoming all in with Cloud Foundry."

CF Summit

Joanne Muszynski, IBM Cloud Foundry Project Manager, gave an update on the upcoming Cloud Foundry Summit, which is only weeks away.

The conference will start off on Monday night, 9th June, with some keynote speakers. On Tuesday there is a full day of keynotes, followed by lightning talks before breaking out after lunch into a business track and a technical track.

On Tuesday evening there are two options: 5:30pm-10pm there is a hackathon and 5:30pm-6:30pm there's an evening birds-of-a-feather unconference. Joanne said there also a third option - going out and enjoying San Francisco!

Same daytime format on Wednesday - keynotes, lightning talks, unconference after lunch.

Joanne reminded us that the event is taking place at the Hilton San Francisco Union Square and that if you book 2 nights at the Hilton you can get a free pass to the conference!

More info about the Cloud Foundry Summit can be found at cfsummit.com.

CF Summit Hack Night

Cornelia Davis from Pivotal gave an update on the hack night that is being scheduled as part of the CF Summit. She said it is no longer going to be a competition, due to the broad range of people who might be attending.

The hack night will kick off as an unconference, where topics will be suggested so that groups can be formed and significant progress can be made around the suggested ideas.

Cornelia said that there may be people there to build their first Ruby app and others may want to get BOSH-lite stood up. Others might have ideas for extensions to Cloud Foundry and want to learn how to structure a pull request the right way.

She encourages engineers from any experience level to come and lend a hand.

Participation is not limited to CF Summit attendees.

As mentioned above, it runs from 5:30pm to 10pm. Cornelia said that they will probably kick off the unconference part around 6pm and will possibly get people demoing things at 9pm. There will be a projector available and hi-res displays.

Cornelia requested that anyone planning to attend registers through schedule on the CF Summit website via the "Schedule This Event" link. That will help with the head-count.

OpenStack Summit

Chris Ferris of IBM summarized the OpenStack Summit as concurrently combining a user conference with a developer design summit. The design portion of the conference was where various project teams got together, worked through their blueprints and planned implementation roadmaps.

Chris, who has been to the past 5 or 6 OpenStack events, says he's seeing a trend towards container management and how Cloud Foundry might serve a suitable complement in this area, providing the PaaS layer to what OpenStack provides at the IaaS layer. Chris did acknowledge some bias in that he obviously was seeking out this kind of conversations.

There were a number of sessions talking about Cloud Foundry and implementation strategies for Cloud Foundry and OpenStack.

Joshua McKenty from Piston Cloud was in a BOSH design session at the OpenStack Summit and shared the etherpad document that resulted from that.

Daniel Krook from IBM ran a session with colleagues Jason Anderson and Animesh Singh entitled "Optimizing OpenStack for Large Scale Cloud Foundry Deployments". He said there was a lot of interest in automation and auto-discovery - anything that will make deployment and collaboration easier between OpenStack and Cloud Foundry. IBM has some work in progress in this area.

Jeff Hobbs from ActiveState pointed to the blog post "OpenStack and Cloud Foundry" which represents ActiveState's thoughts from the Summit. He added that the Cloud Foundry talks were extremely well attended. Jeff said there was a feeling that there was natural affinity between Cloud Foundry and OpenStack which may have had a dampening effect on Solum.

Chris said that he did notice a shift in the discussions around Solum - it's still there, but it seems there is a reconciliation that it still has some relevance in context of OpenStack and Cloud Foundry.

Ferran Rodenas also has a post on Pivotal's blog, "Cloud Foundry at the OpenStack Summit Atlanta 2014".

GlueCon

Cornelia Davis from Pivotal gave her thoughts on Cloud Foundry at GlueCon. While there, she gave a presentation on the elastic runtime and BOSH, and she will post her slides to slideshare soon. Cornelia said that the Cloud Foundry track at GlueCon was unfortunately poorly attended with often no more than 20 people in the room, although there was interest in Cloud Foundry and there is definitely name recognition amongst attendees.

IBM also spoke about BlueMix at GlueCon.

You can see Bernard Golden's "Report from Gluecon: APIs and Docker".

BOSH

James Bayer started off the technical update with a report on BOSH. He said that there are updates for the Go agent stemcells and for Ubuntu 14.04 (Trusty) stemcells. They are trying to move things over to the Go agent from the Ruby agent.

James mentioned David Laing's blog post on how to use AWS spot instances to cut the cost of BOSH deployments. This could result in up to 10x cost savings if you are just doing testing and do not mind waiting a little bit longer to get your instances.

James also mentioned Ferran Rodenas' blog post "Managing Stateful Docker Containers with Cloud Foundry BOSH" which shows you how to use Docker containers with BOSH in a way that you can start up Docker containers that have stateful disks. This differs from other Docker orchestration solutions that assume a state-less disk environment for the Docker containers.

Greg Oehmen from Pivotal said that they are working on release versioning and upgrading so they support Ruby 2.0, 2.1 and 1.9.3. He noted that Ruby 1.9.3 is probably long overdue.

Greg said that the Go agent work is almost done, which has involved getting feature parity between Go and Ruby. He said that Pivotal do not have the Go agent in production yet, but do have it in their pre-production environments. This is a similar situation with Ubuntu, which is going from 10.04 (Lucid) to 14.04 (Trusty). They are starting to publish Trusty stemcells and there are a few issues that they are working. One issue is resolved by upgrading PowerDNS.

Greg mentioned that they hope to leverage some of the work that Ferran Rodenas is doing.

Wayne Seguin from Stark & Wayne asked about the availability of CentOS 6.5 stemcells with the Go agent. Greg said it would be available soon, but the matrix of things to test in CI is getting complex and this is something he hopes they can resolve with community involvement.

Mark Kropf from Pivotal's Runtime team, said that they have merged in some work from Dmitry and done some pre-work for Trusty support in cf-release.

Runtime

Mark said that in the past month refactoring work has been done in the go-router, but the big focus right now is their OpenStack CI environment.

The OpenStack CI will mean that each time there is a new cf-release version we can be sure that it runs on OpenStack as well as AWS.

Work on Application Security Groups will probably kick off in the next week, said Mark. He posted a Google document earlier in May for feedback.

There is some work being done on feature flags in the Cloud Controller, which allows for runtime control over Cloud Controller operations.

Metrics

James Bayer said that the Metrics team is looking to get something demo-able by the CF Summit. He said that they will work off a fork of the Router and thanked IBM for their contributions to the Router.

James said they hope to demo application specific metrics originating from of the router. This would be similar to how access logs originate from the routers. The initial demo will be metrics in the log line, but the long term idea is to have it available via an API.

Services

James informed us that the Services SSO (Single Sign On) dashboard work is ready for us. The MySQL service broker has the canonical example of this, but is limited to only showing how much capacity you have used. So check it out if you have a service that has its own user interface and you want it to have a single-sign-on user experience.

"Service usage events" compliment the application usage statistics, giving an API for service usage. The APIs are just being finished now.

James talked about the work being started on MySQL HA (high availability), which uses MariaDB. This adds multi-node functionality for this service.

CLI

Internationalization within the CLI has been going well, said James. There have been a lot of contributors from IBM working on this out of the Pivotal offices in San Francisco. Klingon is looking good!

Work has been done to get ready for the CLI plugin's implementation.

Performance Acceptance Tests (PAT)

The PAT project is now in the incubator.

Identity Management

James said there has been work done on LDAP and UAA integration, as well as ActiveDirectory. There has been activity on the mailing list around this. The goal here is to simplify things and make it more straight forward.

There is a plan to add LDAP groups, which will be passed along with the OAuth tokens (in the "scopes" section).

There was some work done with OpenStack to integrate Keystone, which James says may have been merged in.

Buildpacks

Buildpacks are moving to being offline by default, said James. This means they can use the assets without Internet connectivity.

The Java buildpack received updates for several Tomcat CVEs and these have been rolled into a new version by Ben Hale.

Ruby and Node.js buildpacks have received significant updates, so that they are up-to-date with Heroku's buildpacks. Offline support has also been added to these buildpacks, which is something that Heroku actually would not care about.

There are brand new buildpacks for Go and Python.

PHP buildpacks are currently being evaluated. Heroku has a PHP buildpack, a Pivotal support engineer has been working on another and Zend is another. At the end of the call, a link to the NTT PHP buildpack was posted by Ken Ojiri.

Diego

Within the last 2 weeks Diego has been able to actually run a subset of applications, said James. He stressed that this is not in production and there is some way to go before we get there. Prior to this, Diego was focused on the staging of the applications, which has been working for the past month or so.

Chris asked about .NET and Docker alternative backends to Diego, since at least .NET was being looked at in the early stages of Diego development. James replied to say that this will not be available on day one, but the Diego team has been working with the IronFoundry team. First they are looking for feature parity with the existing DEAs.

Heroku-style Procfile support for running multiple different processes from the same code-base it is also being looked at.

James said that the Warden root filesystem for the base container used to be very static, but some code has just been merged in to allow this to use the filesystem from a Docker container. This is a platform operator feature that is in the Diego code-base.

Ruben Koster of Stark & Wayne asked if there were plans to support a persistent filesystem in Diego. In response to this James talked about Diego having short running and long running processes and some people might see that as being similar to BOSH. He said that BOSH manages stateful disk and currently Diego does not. In the future there is potential that the same components are used in both BOSH and Diego, which would enable providing persistent filesystem support in Diego.

James said that in his view they might take an incremental step and do something similar to what ActiveState has implemented with ssh-fs, but this would not be done until they have parity.

There was a question about when Diego will be ready for prime-time, to which James responded that we are probably a couple of months away. He also mentioned that it involves retiring certain components, such as health management, since Diego manages this itself.

Tea, mate?

James mentioned an interesting buildpack called tmate which spawns a tmux session within your application's container. This provides a way to access the container for debugging purposes.

If this is used in conjunction with the Heroku multi-buildpack, it can be used alongside any other buildpack and could be quite a powerful debugging solution.

Conclusion

That's all for this week. See in June! If not at the CF Summit, then at the next Cloud Foundry Community Advisory Board Meeting which is scheduled for 25th June at 8am PDT.

Share this post:


ActiveState is one of the founding Gold-level sponsors of the Cloud Foundry foundation of which there are now 33 members, including IBM, HP, Intel, NTT, VMWare, Rackspace and Docker. Our cloud application platform, Stackato, is based on Cloud Foundry and can be installed on any private internal clouds or public and hybrid IaaS platforms. Learn more about Stackato, the leading independent Cloud Foundry based PaaS.

Related Posts

Cloud Foundry Advisory Board Meeting - 2014 April

Impact Update Chris Ferris provided a brief update on the IBM Impact Conference. He noted several BlueMix announcements including updates for the IBM Cloud Marketplace and some additional capabilities for BlueMix. He pointed out that IBM DevOps Services, is essentially an extension of the formerly known "JazzHub" that allowed devops deployments into BlueMix, and “I think potentially into any Cloud Foundry, which is interesting,” remarked Chris.

Read more...

Cloud Foundry Advisory Board Meeting - 2014 March

Diego Progress Alex Suraci from Pivotal gave an update on Diego. They are currently working in AWS and are able to successfully stage applications within their development environment. They hope to move to a larger scale staging environment and then move forward to usage in production. They are looking to 2 orders of magnitude higher scaling, but are hoping to take it even further. Diego Simulator is a project that is used to test Diego's algorithm. Alex said that this is the first area that they are looking at for bottlenecks. The simulators do not actually push any load.

Read more...

Subscribe to ActiveState Blogs by Email

Share this post:

Category: stackato
About the Author: RSS

Phil Whelan has been a software developer at ActiveState since early 2012 and has been involved in many layers of the Stackato product, from the JavaScript-based web console right through to the Cloud Controller API. Phil has been the lead developer on kato, the command-line tool for administering Stackato. Phil's current role is a Technology Evangelist focused on Stackato. You will see Phil regularly on ActiveState's Blog. Prior to coming to ActiveState, Phil worked in London for BBC, helping build the iPlayer, and Cloudera in San Francisco, support Hadoop and HBase. He also spent time in Japan, where he worked for Livedoor.com and met his wife. Phil enjoys working with big data and has built several large-scale data processing applications including real-time search engines, log indexing and a global IP reputation network. You can find Phil on Twitter at @philwhln, where you can ask him any questions about Stackato. Alternatively, email at philw at activestate.com