 So, welcome to Manchester for the first European OpenStack operators meet-up. We've got the OpenStack community across Europe is very well represented here and we've got folks from as far afield as the US and Japan and we've got a packed couple of days of discussion sessions around OpenStack. It's seriously exciting because I believe in this venue and with the 100 or so people who come along we've really captured the spirit of OpenStack and continuing along in the tradition of OpenStack. So, we're going to be sharing best practices, gathering a whole lot of feedback on how people use the software and getting people actively involved. Tell me more about what I can look forward to in the agenda. So, we've got a pretty packed agenda, maybe 30 odd sessions across a couple of days. We've got things happening like the first meeting of the scientific working group which is going to talk about research computing as it relates to OpenStack. We're very fortunate to have the project technical lead of NOVA joining us to hear about all the things that are working perfectly with NOVA and some that perhaps aren't so much. And I believe on Tuesday morning we're fortunate to have the CTO of a company called EnterIT that's just one that's a tender with the European Union to provide cloud there talking about OpenStack in the European context. And tell me about something that is surprising that is coming out of this meeting so far. So, I think there's some unique opportunities for OpenStack emerging in Europe. The concerns about data sovereignty kind of make the European landscape a kind of unique place in terms of opportunity for localised cloud provision. And we're starting to see the results of that now, particularly with federated cloud provision from localised European-based providers. Well, thank you so much for hosting. So, let's talk Keystone. What were some of the hot topics you discussed during this Ops mid-cycle? So, the Keystone group was really focused on the question of Keystone and Federation and for Europe that's a really important topic because you have a number of different organisations, especially the educational organisations that are all very interested in OpenStack or deploying OpenStack and need to communicate with their peers in other countries. So, you have Federation because of the distributed nature of the community here. So, they're very interested in it. When we did sort of a show of hands who's actually using it, there were only three groups that are actually currently using Federation as a group here in the UK that's actually very heavily using it. And they're all interested, but that was sort of the first aspect. The second part of the conversation really then jumped into the rest of Federation. What else needs to be federated within the OpenStack environment? So, we kind of stepped away from Keystone and started talking about what needs to happen in Nova, what needs to happen in Glantz, like if I'm dealing with images and wanting to manage those and manipulate those across different boundaries. And tell us some of the action items that came out of your discussion on Keystone. So, the sort of the top action items were not so much Keystone related as they were. What are the other services that an action need to integrate? I remember there was one group that was looking at an extension to Keystone and I think we've sort of mapped them or given them a path to communicate with the Keystone community, really get involved and actually potentially develop the code that they need to better integrate into the back end service. They're directly based service to tie into the Federation services of Keystone. But beyond that, it was more how do we continue that concept of Federation into the other projects, into the Nova, into the Glantz, into the Neutrons were appropriate. Thank you. Hi Stig, how are you doing today? Very good, very good indeed. Great to hear that. It's a great meet-up, I really enjoy this actually. It's the first time I've been to the Operators meet-up and I've really got a lot of value from meeting so many people. There's just everyone who I've spoken to, everyone I've met has been useful and valuable and helpful. It's a fantastic experience actually, it's really great. Glad to hear that. So tell me a little bit more about the Scientific Work Group. So I work with Cambridge University and we are looking at really sort of rationalising a lot of the infrastructure that my group, the Research Computing Services Group provide. We started looking at OpenStack in the autumn of last year and really I think what we found was that it was very difficult to find other people who were using OpenStack in a similar way to the way that we envisaged. We went to the summit in Tokyo. We started off, we set up ball rolling with a couple of introductions with some really helpful people. Tim Bell from CERN in particular who put us in touch with a guy called David Flanders who is in the OpenStack Foundation now. And really from there we just had this amazing, it was like a Mexican wave going around the world and he put us in touch with other researchers across the UK and then that snowballed across Europe. I went around Australia at least four times. There are so many Australian people who are interested in this idea of this scientific working group came back through the United States and now I think we have something like 85 people signed up for the scientific working group which is just incredible. It's way beyond our expectations. So in terms of a community there is this huge sort of spontaneous upwelling of interest from across the world from these scientific research foundations and institutions. Great to hear that. So what were some of the hot topics you discussed at the Oxford Cycle? So we had a fairly informal get together and the first thing that happened was the room that we had was too small by half and so we had to go and find a bigger space and really I guess the really interesting thing that struck me was like the social connection that was forming as soon as we've raised a topic it was like yes I have this problem or I've solved it in this way or I've seen this interesting work going on over there. There was such a useful exchange of ideas. I mean it was an hour and 20 minutes and every one of those minutes I learned something I think and a very worthwhile experience and I'm really looking forward to continuing with this and building this into something that will be of great benefit to other institutions and other scientific establishments across the world. And so what were some of the actionable items resulting from this session? Well I guess the first thing that we really came away with was this idea that we need to have a way of sharing information. So we already have like HPC tags on the operators mailing list, that kind of thing but I think what we really identified was the need for some kind of a knowledge base like a place for gathering this stuff because a lot of the information that is helpful for doing scientific compute on OpenStack is really scattered around the world and it's not always easy to find or to curate that stuff into a useful collection. So I think that the first and the biggest action item that I saw from our meeting yesterday was how could we actually bring that stuff or link that stuff together into a useful resource which might actually then serve as a knowledge base a planet or a portal, an event stream and an archive of information that will really help people get on board with scientific OpenStack. Great to hear that. Anything else you'd like to add? I think it's early days. This was really... I mean our working group is not even officially inaugurated yet. We're looking forward to doing that at the Austin Summit coming up in April. So I'm tremendously encouraged by this first meeting at the European Operator's Meetup. It's going to be a fantastic journey and I think any other scientific institutions or any other interested parties who'd like to get involved, please do because we are not even officially started yet but already the ball is rolling so please do get started. Great. Thank you. Please tell us a little bit about some of the hot topics here at the mid-cycle. Absolutely. We had lots of great discussions about NOVA and compute-related activities. So I gave a good summary of what was what's been happening over in the TACA and the last few releases with things like Upgrade and all the work on live migrate fixes. So we had a great session going through how we're evolving live migrate for the TACA cycle. We've got some great outcomes there where we're just making sure that all the operators, like their use cases are being dealt with with the work they're working on and that seems to be going really well. We did a deep dive into Upgrades and packaging and everything else. It was really interesting to see the variety of different ways in which people are upgrading and packaging the whole OpenStack and particularly NOVA and how they're dealing with that. Talking about the differences between deploying off-master and deploying from stable branches and how that impacts like the development cycles and book fixing and everything else. So it was really interesting. One of the things that we I got out of this was making sure that the path that we've agreed from previous Ops cycles and other things that we're working out is really having this so particularly things along the lines of the life upgrade work to make sure that we have minimal downtime and hopefully actually zero downtime during the upgrade procedures and seeing how that impacts with people's packaging and their orchestration and what they're currently doing is how we get to more people using it and more feedback. There's some great feedback on folks using the initial versions and how the current roadmap for changes should help them and all that. So tell us a little bit about the action items now that have come out of the mid-cycle. Absolutely. So one great action item that's come out is making sure that we document more of the upgrade procedures so we need to make sure that the stuff that we now have validated in the gate with the Grenade Molteno job to make sure that process gets documented well so people can take that out and scale it out. There are some specific cases about back ports to stable branches and how the stable maintenance team is spinning up to make sure that we integrate well with that and make sure that there's good ways for all the operators to get their bugs and their problems treated in that system. So that's a good action item to follow up with the stable team. There are some more specific things like the step and the resides and the bugs that are closed and the features that we've done to fix individual people's problems to make sure that we ground those folks as actually being completed the way that we want the documentation in place to be more specific. So yeah, it's been really good to make sure that you get the full cycle of the problem sorted. Thank you very much. So Rocky, tell us what is OSOps? OSOps is a user committee project that is operations and it is collecting up all those scripts and tools that operators need beyond the basic OpenStack software to manage the day-to-day operations of their clouds. And tell us about some of the hot topics that were discussed about OSOps. Well, what we did is describing how we actually break down the breaking different repos to make it easier for the operators to start pushing code upstream. This is not the typical code that you will push for other OpenStack projects. This is a little bit more like an open space for operators to push whatever they want. So we're walking through the process to actually which one of these repos is the most important for the code that they wanted to push upstream to use those scripts, to use that code to sit back. One of the good things about it is we're going to start another session in this meetup to provide a little bit more like walking in, what exactly you will be welcome to push in those repos. So tell us about some of the action items that we can look forward to coming out of this summit. Well, one of the items that came out of the discussion was the need for tools that monitor the actual health state of the individual software packages in the cloud as opposed to just the infrastructure hardware. And the ask was how are you doing this? We've already got some info from CERN and we're going to try and get that merged in so other folks can take a look and give it a spin. And I think that we also need to provide a little bit more organization in the future. It could actually look a little bit messy and we're going to take it right now because it's just an initiative right now. But as we start making more content available for operators we're going to create a little bit more structure around it and we're going to provide the proper documentation for everybody to start documenting even more what these OS ops project is about. Right and one of the things that I've taken away from this is that we need to get also Sphinx documentation format into our repos so that each of these scripts do get documented and there's an easy way for operators to do that. Thank you very much for your time. Thank you. So what would you say are the top bugs identified during the mid-cycle? There's a lot of great discussion about improvements needed for other services to support Keystone v3. There was discussion about neutron support for IPv6. There was a lot of talk around the different SDKs and OpenStacks CLI improvements needed for user tooling and developer tooling. So I think those were some of the highlights in the discussion. Great and tell us also about the highlights in terms of new feature requests. Yeah, there's ongoing work around a topic called get me a network which is essentially to make it easier to provision your first VM network is immediately available. Talk about improvements and kind of next steps needed there. Some other features were actually given the operators focus were tooling that the operators were interested in. It's really interesting within the meeting other operators mentioned that they had contributed stuff to OS ops tooling repo that were available for those other operators who were wondering if they should go build something. So a really great example of collaboration between operators. They're just that we're taking the feedback from this meeting and delivering it to the different project teams who weren't there in the meeting to kind of say these are the things that operators are really focused on us improving within OpenStack and hopefully those become a priority going forward. Thank you. Thanks. So tell me what were some of the hot topics that you discussed today? So with the Neutron project the number one topic that we ran into I think was a real concern about documentation. There were a number of different users that were basically saying that they didn't have enough documentation to really clearly understand what the best practices were which resources to actually use like which virtual switch to actually leverage and then proper configuration for those and so we actually pulled together a pretty decent list of a couple of additional scenarios that were needed to be captured in the documentation that I think would really help them. Great. So Edgar tell me about what some of the items that came out of the discussions were. Yes. As Robert mentioned, documentation was very important. So one of the things that we highlighted in the session was that we have a very specific guide for networking. We describe the guide now as a typical documentation when you go step by step. So actually what we did in that one was to create a scenario. Typical deployments that the operators would actually ended up doing to guide them to have best practices on how to use Neutron and how to actually get the best practices out of it. One of the things that we missed in that networking guide is a good section about how to do a troubleshooting at the booking on the communication path between virtual machines and the physical host. What else do you want to tell us about what's been happening at the mid-cycle? The mid-cycle I think has been great. We're getting a lot of feedback from the operators. A lot of interesting interactions especially given that many of them are running systems at very large scale. Here in Europe it seems like it's very high-performance computing so it's a large scale computing users. The network is an important part of that. Questions about network performance came up. How can we improve this? How can we find better ways of doing this? As Edgar was saying, one of the biggest issues was troubleshooting. When something does go wrong, how do they determine that? Just getting that feedback I think was really useful to get back to the community too. So besides documentation, what feedback would we be taking to the Neutron development team? Operators are aware that Neutron is a very complex architecture. They want actually to simplify a little bit based on the use cases that they have. They don't want to deploy all the simple use cases that we have available in Neutron. On the other hand, they want to have whatever is useful for the use cases to be applicable through a more simplified architecture. Thank you. Tell me your perspective on being here at your first Ops mid-cycle. It's unique to actually get here and be with other operators. It's nice to get out of the U.S. and get into the European zone. You get different perspectives on how they solve problems, the types of problems they have. It's very academic here it seems like, and the integrations that they approach. Whereas we are very industry focused, and they prefer peer-to-peer approaches that can bounce from cloud and do their own integrations, which is very core to their problem solving abilities and what they need in their final solutions. Tell me about what you're seeing here at the Ops Summit. So, what I find really exciting is creating a community of operators where they can talk with one another and exchange ideas and maybe find some commonality in some of the issues that they're finding so they can come to resolution. And what surprised you so far? I think what surprised me as well is some of the differences that we're finding in the European Union versus the United States and maybe North America in terms of how some of the requirements that are being placed on the operators particularly during the bidding process. What are you learning that might be helpful to your UX work in OpenStack? Well, it's actually a huge amount of really rich data that we're driving from these operator summits, of course, right? So now we have all these EtherPads and that's a great opportunity for us to go through there and start coding the information and trying to come, I guess, create like larger themes based on what we're seeing here. What else would you like to add? So, I think one of the things that we do need to do is so we create these EtherPads it's a lot of really rich data but somewhat unstructured. I think one of the things that we can do is start taking that data and I worry that it just sits there and it doesn't do anything and what I'd love to be able to do is as a group sort of say this is what we saw these are the larger themes so we have something to march on. Hello. Thank you for joining us. So what's your take on the operators this week? Well, it's a first for me. It's my first motorcycle meetup so it's interesting to see what happens and to meet other people that might have the same problems the same success also because there's a lot of positive things in the OpenStack community it's really a chance to exchange with a lot of people who run various types of clouds, different usage different sizes so it's interesting and I think it can make things move forward it can help discuss with the developer community and just not the operators inside and the developers on the other side so I think there's a lot of positive things that can come out of this meetup if we can follow up because it's not that you can make everything go much better in OpenStack so I guess it's just a starting point but don't together so it's a pretty positive thing. Thank you so much. Hi, thank you for joining us today I was wondering, could you please share your experience with the Ops mid-cycle in Europe? I think it's a very important meeting for two main reasons one is the EU it's a European-based meeting so Europe has been doing some giant steps forward we enter as our company and the cloud team alliance has been awarded for a EU tender, very important because the EU institutions will be using OpenStack for public services and the other very important reason is because it's very focused on operators operators are that group OpenStack users that cope every day with real life problems and with customers so they are the connecting point between the work of developers and the needs of the users that's why it's so important for us to be here You're welcome. Hi, please share your experience with the European operators mid-cycle. The mid-cycle has been great there have been a lot of really good conversations but in session and out of session the venue has been fantastic we've been able to have just one-on-one get-togethers as well as larger groups both in and out of session as well. So what were some of the key takeaways from the sessions here? How people are operating seems to be very different from something we're used to in the US the regulations seem to call for different policies different procedures somewhat of a different game plan and it seems as more companies attending are closer to Trump which is interesting it's not always the experience that we've had Thank you so much.