 Well, hello, and welcome to the Newton Design Series. I'm Heidi Joy-Trathaway with the OpenStack Foundation, and I'm here with Sean McGinnis. Would you please introduce yourself and tell us a little bit about you? Sure, I'm Sean McGinnis. I'm the current project team lead for the Cinder Projects for Newton. I've been working on OpenStack now since Ice House, and I work for Dell in our Minneapolis office as a senior principal software engineer. Great. Well, tell us a little bit more about Cinder. What does it do? Cinder is the OpenStack project for managing block storage devices. It provides an abstract interface over the top of various traditional block storage backends to be able to use things like LVM or a vendor separate physical array to provide block storage to virtual machines and containers. And Cinder is one of our most adopted projects in OpenStack, one of the top six most adopted projects. Tell us about some of the hot topics that your team discussed in Austin related to Cinder. Sure. Well, a lot of the discussion was actually probably not too interesting for end users directly. We had a lot of changes in Mataka that effect. Different things that the developers working on the code need to be aware of. Things like now implementing microversions in the API, doing things like rolling upgrades. So there's a lot that we need to be aware of and then hopefully these features as a matured will actually be useful to the end user. Things like the microversions will allow a consumer of the OpenStack cloud to request a specific version of the API. So we're able to keep going and making changes but they're able to get the behavior out of the API that they expect. Other things for replication. Again, that's been an ongoing theme. We are continuing to evolve that and then just things like user messaging on errors and various issues. Great. Well, then tell us more about your Newton planning and specifically the kinds of user needs that your team identified that you're working to solve. Sure. Since I brought it up, I guess I'll talk about replication first. We've had some development over the last couple of releases for replication. We have replication support. They're not adopted by all vendors that are able to do replication. So one of our priorities is to expand that coverage. Hopefully we'll get, I think we have five or six back-end vendors right now that support it and we'd like to expand that to have it be a more common functionality, not just a niche thing. There's a lot of talk about expanding on replication and really until we get that good adoption, I'm a little hesitant to add more functionality on top of it. That said, a lot of this is really driven by our own conversations. I'd love to hear any feedback from end users up there, what they're looking for with replication, what their needs are so that we can make sure that we meet their use cases. The other, another top priority for me, at least what I'd really like to push in the team, we have a few people working on user messaging. That's the idea that being able to return internal error reasons better back to the end user than what we have right now. Unfortunately, a lot of times the experience isn't so good when things go wrong. If you're using Horizon and you create a volume, you might just get back a really generic message saying that it failed. And then it's up to digging through log files and repping messages and things to find out. We'd like to make that experience better for end users, be able to actually tell them this failed and here's why. And then another hot topic and priority was kind of a general theme, I think, for all the projects right now is moving over to using the OpenStack client. As it has been so far, we have a Cinder client, that command line interface, and it works great, it has all the functionality, but it's specific for Cinder and there's one specific for Nova and there's one specific for all the other ones. So in case anyone hasn't heard, OpenStack client is the one CLI to rule them all. It's pulling in all that functionality from all the different clients. So we want to move to that. We do want a one consistent good user experience, but there's some functionality that's in the Cinder client that we haven't gotten into OpenStack yet. So there's been talk about deprecating the individual projects CLIs. And I can't go forward with that until I'm sure that we're not going to lose any functionality by directing everyone to the OpenStack client. So we do have a couple of people looking at that, seeing what's missing and getting those two up to par so that really if we do get rid of the CLI from Cinder, that end user can still do everything that they need to do through the OpenStack client. I guess one tricky part of that we needed to figure out was that we just recently added a plugin for our CLI for our Brick client. So OS Brick is a library that handles all of the storage interaction on the operating system for being able to do things like re-scan and discover new SCSI devices when we expose a new volume to a host like a Nova compute host. So we added a plugin to Cinder CLI to be able to do that outside of Nova for being able to do things like ironic or really just storage management directed by the end user just on their own. So we just had to make sure that that functionality can still be met through OpenStack client. Looks like we might have a path forward there. So hopefully we should be able to move on to OpenStack client and have that one common interface for everything. Sounds like a lot of new great features and enhancements. As you're looking ahead toward Newton's release in October, what would you say are the top three new features or enhancements that your team plans to deliver? Well, I think another recurring theme, active-active high availability, that's some progress was done in Metacor for that and there's still some pieces that need to get put in place, but I'm hoping it looks like we might be able to get that in Newton. That'll be great for making just the services more robust. We're looking at replication, as I mentioned, really would like to see more support there. Test coverage is actually a priority for us. There's good coverage of the core functionality of Cinder, but there definitely are areas that have very light or really no test coverage. And we wanna make sure all of that is being covered through CI tests, functional tests, unit tests, just to make sure that whatever we're delivering when end users pick it up, there's no surprises. Great, so one more question for you. As you know, the product work group has established some themes that we use to connect the dots between four projects, and those themes are scalability, resiliency, manageability, modularity, and interoperability. So I wanted to know which of those themes do you intend to address most heavily in Newton? Of course, all of them to some degree, but I guess I would say the top two, if I picked two out of there would be resiliency with things like high availability. We want Cinder to be solid and manageability with things like the error reporting back to the end user and making things like open-side clients accessible, usable by everybody. Great, is there anything else you wanna add to the book swatching? I guess just always asking anyone interested in getting involved, we can always use help, code reviews, documentation, really anything. You don't need to be a programmer. If you wanna be part of the community, get involved, we'd love to have you. You can stop in on the IRC channel, open-stack Cinder, and just ask if anyone needs any help, and I'm sure some people would have some suggestions. Thanks, Sean, for telling us about Cinder. Thank you.