View from 30,000ft of Some SUSECON Ceph Sessions
As I sit here at 30k feet on my way to SUSECON in Nashville, I am excited about the future of SUSE and the solutions we are able to craft with partners and for our customers to consume. I’ll be talking about a few of these solutions this week during the five sessions I am presenting this week and will try to give you a quick preview here. If you’d like to join one of my sessions, you can find them here: https://www.susecon.com/sessions.html?search=byte#/
First, is a session covering architecting Ceph based solutions. I’ll talk about technologies and how they impact design when taking the use cases the cluster is being used to address into account. I’ll also be briefly covering the broad list of guides we have available and what it takes to build one of these guides.
The second session is a hands-on lab where Darren Soothill, the Sr. Storage guy in Europe, and I will be working with the attendees to understand device technologies and the implications of those in building a solution. We’ll also be taking the attendees through the process of architecting a few cluster solutions and hopefully shine some light on some of the mystery that is storage architecture along the way.
The third session is the result of a long term work effort around Veeam between myself and Alex Zacharow from our ISV certification team. I’ll be co-presenting with Steve Firmes from Veeam and we’ll briefly cover Veeam architecture and options and then talk specifics about implementing it with SUSE Enterprise Storage and the options available to you there. I’ll also discuss a few lessons learned along the way concerning tuning storage for VMware servers and optimizing the Veeam proxy.
Session four is all about spinning your propeller with performance tuning and numbers for all flash Ceph clusters. I’ve had the privilege of working with Intel on hammering out a lot of IOPS with a 10 node, 120 OSD cluster. In fact, I’ve blown past 1 million IOPS without pulling the tricks of disabling logging, authentication, etc. that we see so commonly when people are talking about their “high performance” Ceph clusters. As part of the testing, I’ve also been able to do a little analysis of using Intel Optane PCIe based storage and have some thoughts on how to use it in real-world cluster environments.
The fifth, and final session that I’ll be presenting is a co-presented session with Kishore Gagrani from DellEMC. We’ll be talking about the work done on the recently released implementation guide for SUSE Enterprise Storage on DellEMC PowerEdge servers. As part of the discussion, we’ll outline what we did, and perhaps more interestingly, what some options are that we are exploring for joint work in the future.
As you can see, I will be quite busy, but if you see me taking a breather, please feel free to catch me to visit about SUSE Enterprise Storage. I’d love to hear how it is helping your business or answer questions you may have about how to implement it to solve your challenges.